lightning protection grounding solutions for communication guide.pdf · lightning protection &...

Lightning Protection & Grounding Solutions for Communication Sites

LightningProtection &

GroundingSolutions for

CommunicationSites

FIRST EDITIONKen R. Rand

Published by PolyPhaser ®


© 2000 by PolyPhaser®

All rights reserved. No part of this publicationmay be reproduced or transmitted in anyform or by any means, electronic ormechanical, including photocopying,recording, or any information storage andretrieval system, without permission inwriting from the publisher.

First Printing 2000

Printed in the United States of America


IntroductionThis publication was compiled from the original book The “Grounds” for Lightning and EMP Protection byRoger R. Block, co-founder of PolyPhaser Corporation, and additional articles written by Roger and myselfover the last several years. I have brought in some up-to-date information during the re-write. Some text hasbeen revised and re-ordered for logical sequence and clarity.

The lightning protection industry owes a great deal to Roger for an innovative lightning protector productline, site protection techniques, and his informal way of pointing out problems. Although seldom appreciatedby traditionalists and the academia, his research and practical conclusions were “right on” for our emergingindustry. The following chapters continue in that tradition.

Thanks to all who participated in this latest effort and an honorable mention for Bogdan (Bogey) Klobassawho contributed to Chapter 7 and never tires of talking about lightning protection.

“Let us hope this book, as with our knowledge, will never have an end, for we are just beginning to learn andunderstand.” 1

Ken RandJanuary 2000

1 Quote from Roger Block, who is now happily flying his airplane, inventing something, or reluctantly playinggolf.


CONTENTS

Table of Contents 1 The Lightning “Event” ...........................................................................................................1

2 Tower Strikes & Solutions .................................................................................................11

3 Grounding & Materials .......................................................................................................27

4 Ground Impedance ..............................................................................................................39

5 Tower Top, Pole-Mounted & High-Rise Communication Sites ....................................43

6 Coaxial Cable Lightning Protectors ...................................................................................47

7 ac and dc Power Protection at Communication Sites .....................................................55

8 Telephone Network & Computer Interfaces at Communication Sites .......................65

9 Protecting Equipment from NEMP Damage....................................................................69

10 Security Cameras, CATV, GPS & Satellite Protection ....................................................71

11 Appendix ...............................................................................................................................77Lightning Protection Myths and Traditions ............................79Tables and Formulas ..............................................................83Definition of Terms ................................................................85Bibliography ...........................................................................88

Lightning Protection & Grounding Solutions for Communication Sites 1

CHAPTER 1

The Lightning “Event”here are volumes of information availableon what we believe lightning is and howwe think it works, most of it beyond thescope of this modest textbook. We willindulge in a form of pragmatism focusing

on a practical approach to equipment protectionat a communications site during a lightning “event.”The science of grounding (earthing) for lightning

events encompasses both the laws ofphysics and RF design.

Throughout this textbook areproven concepts, which willprotect your valuableequipment from direct orinduced lightning damage.Whether your equipmentis at radio site, pipe

line, utility sub-station,telephone central office,

maritime, military, or sensitivesecurity installation, the same

requirements apply for protection devices, properdevice placement, and earth grounding.

THE STEPPED LEADER AND THEUPWARD GOING STREAMER

As the electrically active cloud stratifies its chargein preparation for a cloud to ground strike, itproduces an opposite polarity “mirror image” areain the earth directly below. Most cloud to earthstrikes are negative (electron flow downward),some are positive, and an occasional event is bi-polar. Positive strikes are usually more severe andhave been associated with cyclone activity(tornadoes/hurricanes). To keep things consistentthroughout this book we will be using negativestrikes in our examples.

As the “E” Field (voltage) builds in potentialbetween the charge center in the cloud and

earth, it reaches a state where the atmospherebegins to break down and a “stepped leader” fromthe cloud tentatively reaches out and downtowards the earth. Although the stepped leader isalmost invisible, it is forming the beginnings of anionized path that the strike(s) will follow on its wayto an upward going streamer (also known as a“return stroke”) or direct earth contact. Thestepped leader jumping distance is determined bythe charge in the cloud. The smaller the charge,the smaller the jump. A typical jump (96%) is 150feet or greater. The stepped leader will move thisdistance in 1 microsecond, pause for 49microseconds, and then make another jump.

As the end of the stepped leader (which has thesame potential as the charge center in the cloud)approaches the earth, the “E” Field gradientbetween the end of the step leader and any high“earthed” conductor (trees, towers, “lightningrods!”) exceeds the breakdown of the atmospherearound the “earthed” conductors. A corona formsaround the part of the conductor closest to theincoming stepped leader. If the stepped leaderapproaches closer, the corona grows in to whatwe call an “upward going streamer” representingthe opposite charge in the earth. This streamercan reach out 15 to 20 feet in an attempt to joinwith the stepped leader to form a conductive pathfor the main series of strikes to follow.

Once the stepped leader and streamer are joined,large currents will flow as a consequence of thehigh potentials involved. The amount of currentflow in each stroke is determined by the ability ofthe cloud to migrate more electrons to thedischarge point, and the overall inductance of theionized path and struck object. This entirediscussion is applicable only until a newer andbetter theory comes along!

T

Lightning Protection & Grounding Solutions for Communication Sites2

STEP LEADER IMPLICATIONSThe “Rolling Ball” Theory

If the tower is over 150 feet tall, side-mountedantennas are vulnerable to direct hits. Since 1980,the NFPA (National Fire Protection Association) hasbeen advocating in their Lightning Protection CodeNFPA #780, that a 45-degree cone angle from thetop of the tower towards the earth does notdescribe an effective protection area.

Visualize a tower site, and imagine a 150-foot radiussphere (representing a step leader typical jump)rolling over all outlined objects, everywhere thesphere touches could be hit by lightning. Thesphere must be “rolled” for each compass linesince we are dealing with a three dimensionalimage. When the sphere bridges between twopoints, the area beneath the sphere is a 96%protected zone.

96%Protected Area

Ø 300'

150'

150'

96%Protected Area

150'

150'

Ø 300'


As the sphere rolls up the tower, it will begin totouch side mounted antennas above the 150-footmark. For guyed towers, the sphere will need tobe rolled not only for each compass line aroundthe tower base, but also around each compassline for each guy anchor point. The mesh that iscreated will cover the tower like canvas on a circustent. The area above the tent is unprotected andthe area below is the protected area.

Side-mounted antennas near the top, or in sectionsnot covered (protected) by the guy wires, can behit. One way to protect these antennas is to installtwo or more horizontally mounted “lightning rods”attached to the tower just above and below theantenna. As the 150-foot radius sphere rolls onthe tower, the length of the horizontally mountedrods protrude outward from the tower so thesphere does not touch the antenna.

For a 20-foot long antenna, side-mounted abovethe 150-foot height, the horizontal rod(s) shouldprotrude a minimum of 6 inches beyond theantenna. This will give a 96% degree of protectionfrom direct strikes to the side-mounted antenna.Since diverter rods are horizontal and are locatedin the end nulls of the antenna pattern, no changeswill be made in the systems performance.

The rolling ball concept is based on the step leaderjumping distance. The larger the charge in thecloud, the larger the jumping distance. The smallerthe charge, the smaller the distance. This is whythe percentage of protection for the zone (96%)is not 100%. Theoretically a small step leader couldpenetrate the zone, but it would be a small strikewith little damage capability.

A tall tower, above the 150 foot point, should havecoax cable grounding kits spaced so a side striketo the tower will not have to go far before a bondbetween the tower and transmission line(s) occur.This will help prevent side flashes, which couldproduce water invading pin holes in lines. Arecommendation is for 75' to no more than 100'separation between grounding kits above the 150'point- unless the rolling ball concept shows guyline protection.

ProtectedArea

ProtectedArea

Non-ProtectedArea

Non-ProtectedArea

150 ft. Raduis Circle

Horizontal“Lightning Rods”

150ft.

Tower

Not ProtectedHeight

Protected

Not Protected

50ft.

100ft.

150ft.

200ft.

375ft.

300ft.

225ft.

150ft.

GroundingKit forShield ofCoax


STROKES AND STRIKES

One IEEE Standard is an 8/20µs, 3kA currentwaveshape for lightning (see Chapter 7 forwaveshape and discussion). This is the waveshapeexpected to occur at the equipment after the seriesinductance of the tower and interconnectingconductors rolls off the fast rise time (conservingsome of the rise time energy in the resultingmagnetic field), and reinserts the conserved energyat the end of the stroke, affecting the pulse decaytime. This standard was originally for ac powerapplications and has been carried over to coaxialcable entry expectations. With today’s heavilyloaded towers and multiple coax runs to theequipment, one can expect a much faster rise timeand larger current flows.

Lightning typically takes the form of a current pulsewith a very fast rise time. Recent studies haveshown that lightning pulse parameters can varygeographically. The measurement test setup andthe inductance of the struck conductor can alsoaffect results. The pulse statistics in this book arefor illustrative purposes showing the kinds of pulsesthat could occur and were taken from a series ofmeasurements done in the U.S. during the 1970’s.

A typical strike (in this series of measurements)could have a 2µs rise time to 90% of peak currentand a 10 - 45µs decay to 50% of peak current. Thepeak current will average 18kA for the first impulse(stroke) and less (about half) for the second andthird impulses. Three strokes is the average perlightning event.

A strike is a constant current source. Once ionizationoccurs, the air becomes a conductive plasmareaching 60,000 degrees F and is luminous. Thisluminosity level is brighter than the surface of thesun! The resistance of a struck object is of smallconsequence, except for the power dissipation onthat object (I2 x R). Fifty percent of all strikes willhave a first strike of at least 18kA, ten percent willexceed a 65kA level and only one percent will haveover 140kA. The largest strike ever recorded wasalmost 400kA.

DISTRIBUTION OF TIME TO PEAK CURRENT

DISTRIBUTION OF TIME TO HALF CURRENT

DISTRIBUTION OF STROKE CURRENT - kA

0.01 0.1 1 2 5 10 20 30 40 506 0 70 80 90 95 98 99 99.9 99.99.01

1

10

100

12µs

% > ORDINATE

TIM

E T

O P

EA

K C

UR

RE

NT

- µS

5.8µs

1.8µs

0.66µs

0.25µs

.001 .01 1 2 5 10 20 30 40 506 0 70 80 90 95 98 99 99.9 99.991

1 0

100

1000

170µs

% > ORDINATE

TIM

E T

O H

ALF

CU

RR

EN

T -

µS 100µs

45µs

17µs

10.5µs

1.5kA

.001 .01 1 2 5 1 0 20 30 4 0 50 6 0 70 80 9 0 95 98 99 99.9 99.991

1 0

100

1000

140kA

65kA

18kA

6.2kA

3.1kA

% > ORDINATE

PE

AK

CU

RR

EN

T -

kA

70kA

32kA

9kA

3.1kA

First Return Stroke

SubsequentReturn Strokes


WHY TOWER SITES AREDAMAGED

Tower sites are struck by lightning more often thanany other site. The reason is obvious; the toweris higher than the surrounding terrain, and it is aconductor! Tower structures have a certain amountof resistance and inductance per foot. Most peoplethink of resistance when talking about lightning.However, a tower with all of its weight has rathersmall joint resistance, typically less than .001 ohms.The E=IR drops are considerable when 18kA istraversing, but even larger peak voltages arepresent during a lightning strike.

DISTRIBUTION OF THE NUMBER OFRETURN STROKES/FLASH

Chart shows susceptibility versus heightbased on Westinghouse data.

Inductance for either coaxial lines or singleconductor grounding wire can be estimated byusing the tables below.

Every conductor has inductance. The amount oftower inductance is dependent upon its geometricconfiguration. The width-to-height ratio willdetermine the total inductance of a tower. Atheoretical self supporting 150 foot tower, with a35-inch side width, can have an inductance of about40µH. This value of inductance can beapproximated (W/H < 1%) by treating the toweras a 1/4 wave antenna using:

468 x 106

2 (H in feet)= f

then the inductance L =377

2πf

Coax Diameter

1/2” 7/8” 1-1/4” 1-5/8” 2” 3”

51.0 48.0 45.7 44.2 43.0 40.4

81.0 76.0 72.3 70.0 68.0 64.3

111.0 104.0 100.0 97.0 94.2 89.2

174.0 164.0 157.3 152.5 148.7 141.2

306.0 289.0 277.8 270.0 263.4 251.0

100

150

200

300

500

Approximate Inductance in Microhenriesfor Coaxial Lines

Len

gth

in F

eet

Approximate Inductance in Microhenriesfor Conductors

Size and (Diameter)(0.46) (0.365) (0.257) (0.162)(0.102) (0.064)

Strap 6” 3” 1-1/2” 0000 00 #2 #6 #10 #14

1.07 1.28 1.49 1.68 1.75 1.86 2.00 2.14 2.28

2.56 2.98 3.39 3.78 3.922 4.13 4.40 4.70 4.98

4.21 4.83 5.46 6.04 6.25 6.57 7.00 7.42 7.85

5.96 6.80 7.63 8.40 8.70 9.10 9.70 10.25 10.81

7.78 8.83 9.88 10.85 11.20 11.70 12.44 13.15 13.85

9.67 10.93 12.19 13.35 13.78 14.40 15.26 16.11 16.96

5

10

15

20

25

30

Len

gth

in F

eet

2 - 3 Strokes

% > ORDINATE

NU

MB

ER

OF

RE

TU

RN

ST

RO

KE

S /

FLA

SH

5 - 6 Strokes

10 - 11 Strokes

0

4

8

12

16

20

24

200 400 600 800 1000 1200 1400

nu

mb

er o

f h

its

per

yea

r

height of object in feet


The divided voltage present on the coax shieldcreates current flow through all the additional pathsto ground attached to it.

GUYED TOWERS

We have looked at a self supported tower and canreasonably conclude that, without properprotection and grounding, our equipment will sufferdamage.

Looking at the current distribution on a guyedtower, we see the guy wires and grounded guyanchor points perform an important role during alightning strike.

The same 150 foot tower, with 35” side widths,will be used as the example. The use of ½”diameter guy wire with no insulators would looklike the following drawing.

Consider a ½-inch diameter coax running down 135feet from the top of our theoretical 150 foot tower.It will have an inductance of about 72µH. If thecoax shield is grounded at the top, as it should be,and at the 15 foot level of the tower (a locationthat we shall see is not optimal), then the totalinductance of the tower would be:

If the coax line is pulled away from the tower atthe 15 foot level, traverses 20-feet horizontally tothe equipment building and goes to a ground barhaving a 6-foot long, #6 ground wire, the totalshield inductance for this path is 12.7µH. Toaccount for each directional change, one for thecoax bend at the tower and one for the groundplate, 1mH was added. This figure is used tofacilitate calculations. The real value for a sharpbend is more in the order of 0.15µH and isdependent on the size and shape of the conductor.

If a perfect conducting ground system (with a non-inductive connection) were present, a 2ms risetime, 18kA, constant current strike, hitting thetower would develop an -L di/dt drop of 243kVbetween the top of the tower and the bottom.The height at which the lower coaxial cable shieldkit is bonded to the tower and pulled away fromthe tower toward the equipment determines thevoltage that is present on the coax shield.

A 180.28 99.00 33.00B 141.42 75.60 25.20C 111.80 58.14 19.38

InductanceLength Inductance (µH) for 3

Example in Feet in µH in Paral lel

72µHCoax

36µH Tower15' to 150'

4µH Tower10' to 15'

24µH

4µH

4µH1µH

2.4µH

24µH1µH 8.3µH 24µH

12.7µH4µH 3µH

24µH

27µH

150'

B

C

A


On a triangle base tower, where “A” is approximately180 feet long or about 99µH each, there wouldbe 3 “A’s” in parallel or 33µH total inductance. Thiswill significantly change some of our L di/dt values!Likewise, the lengths of “B” and “C” would be usedto calculate their inductance contributions. Thething to remember is - “B” and “C” touch the maininductor (the tower) at different heights(inductance). These heights must be transformedinto their appropriate values of inductance beforethe values of guy inductances can be combined.

To keep it simple, our guy attach heights are at150 feet, 100 feet and 50 feet. Our completestructure looks like this:

Re-drawing the tower circuit:

Resulting calculation equals about 12µH top tobottom.

When the 18kA lightning strike occurs, it will havea voltage drop of

-E = L di/dt = 12µH x18,000 = 108kV 2µs

from top to bottom ground.

This is less than half of the voltage drop of the selfsupport tower without guys.

The distribution of current on this set-up is a littlemore complicated. Using mesh current networkanalysis:

Average Coax Current is 2.79kA

The coaxial cable run to the ground outside barwould have only 1.26kA going to it and would beelevated to 2.14kV. Again, this is far less than the4.3kA and 7.3kV of the self-support tower!

Before you pull down your self-support tower,remember, in our example we kept the sametower side width of 35 inches and just added guys.A guyed tower might not be this wide, but wewanted to point out the improvement that the guysmake by using the same size tower in ourcalculations.

All of the previous calculations assume the guysare without insulators and the guy anchors arebonded together with the tower leg grounds toform one ground system. If this is not done, theground resistance/surge impedance at each guyanchor would determine the current distribution.

72µHCoax

3µHTOWER + COAX + BULKHEADPANEL + GROUND WIRE

9.3µHTower

13.3µHTower

13.3µHTower

150'

19.38µH

33µH

AGuys

BGuys

CGuys

25.2µH

100'

50'

15'

72µH

33µH

13.33µH 13.33µH 9.3µH

25.2µH 19.38µH 3µH

Coax2.79kA

6.58kA

8.63kA 13.33µH150' 100' 50'

9.3µH15'

4.02kA 2.05kA 5.33kA

18kA 7.995kV

Guys1st Set 2nd Set

Guys3rd SetGuys

Tower + Coax + LinesLast 15' of


Now that we have the current distribution, let’ssee what happens if we ground the coax shield;not only to the bottom and top, but also groundthe coax at the guy attachment points on the tower.The new circuit would be:

So the total current distribution is:

Average Proportional Coax Current is 2.733kA.

Any additional grounding of the coax, say to everytower section, would not provide any benefit forthis size tower (150 feet and less). However, it isimportant to ground the coax lines more oftenwhen above this 150-foot level. The guy wire pathsto ground give the reduction in current on the coax.

A comparison of the two examples shows that thegrounding of the coax at each guy location will givea higher coax current between the 150-foot to 100-foot levels. Here it is increased 39% over the“bottom only” grounding situation. What if wedidn’t ground it at the 100-foot level, but kept the150-foot, 50-foot and 15-foot locations grounded?

Average Proportional Current is 2.775kA.

The coax currents are somewhere in-between thelevels of “grounding at each guy location” and“grounding at the 15-foot level only”.

If we look at the average coax current, we have amaximum 2.79kA for the single ground at 15 feetand a minimum of 2.733kA for the multi-guygrounding. Note the voltage at the 15-foot levelon each example. They do not vary more than about8%. This is a very small reduction for the amountof effort and cost involved in the additionalgrounding installation.

MUTUAL COUPLING

Mutual coupling is the name given to the linkageof the magnetic lines of flux between oneconductor and another. In most cases, it isdescribed using two non-ferrous (non-magnetic)conductors (copper, not steel). However, in ourapplications, we have one of each. The tower (steel)will cause the lines of flux to be concentrated inclose proximity. We also need to take into accountthat each tower leg will share (divide) the currentpassing through the tower. A coax running downone leg would not have a very large coefficient ofcoupling of flux lines, even with the steelconcentration. We estimated this coefficient to be0.166.

Using the formula:

M = k eL1 L

2

where k is 0.166 and L1 and L

2 are the tower and

coax inductances, respectively.

33µH

13.33µH 13.33µH 9.3µH

25.2µH 19.38µH 3µH

26.66µH 18.66µH26.66µH

18kA 7.35kV

Coax Coax Coax3.85kA 2.39kA 1.63kA

3.85kATower

6.58kA

4.84kATower

3.27kATower

4.34kA 2.33kA 4.90kA

At 15'Tower Level

18kA 7.545kV

3.16kA 1.68kA

8.37kA

6.47kA

4.27kA 3.35kA

4.10kA 2.4kA 5.03kA

At 15'Tower Level


In the self-supporting tower where the tower had40µH and the run of coax was about 72µH, Mwould be 8.9µH. This is a significant amount ofadditional inductance. At 18kA, our strike currentand 2 microseconds rise time, this is an L di/dt of80.2kV or a 33% increase!

Additional worst case consideration might be givento the possibility of a low inductance self supportingstructure with a single coax running down the side.Depending on how the coax was attached, if thestructure was tall (> 150 feet), and the coax shieldwas grounded to the structure at the top andbottom only, there would be a large difference ininductance between the two paths. Magnetic fieldcoupling (k) between the two paths would createa reverse EMF on the coax, opposing thedownward energy flow. At some point,approximately in the middle of the structure, therecould be a high peak voltage differential betweenthe coax shield and the structure. This high peakvoltage differential could arc through the coax PVCouter jacket to the structure, damaging the coaxshield. Additional grounding kits could solve thisproblem.

In the guyed tower, the coefficient of couplingwould be the same. But since there is less totalinductance with current flow on the guys, therewill be less current on the coax, making the dv/dtless dramatic. The grounding of the coax shieldalong the tower will segment the amounts ofmutual inductance. The mutual coupled inductancewill then add about 7% to all inductances andvoltages we have calculated on all combinationsof coax shield grounding.

So far, we have taken a look at the currentdistribution on two theoretical towers for a typicalstrike. What happens to the coax line and theconnected equipment in the building when thispotential is present?

IT’S WRONG!

If we look at where the coax leaves the tower onits way to the equipment building, we see thetower will carry the major part of the surge to earth.The outside master ground bar will have 4.3kAdelivered to it by the coax and be elevated to 7.3kVabove earth ground. The master “ground bar” isno longer a ground, but instead a source forelevated potential to be transferred to whateverequipment is connected to it! The above currentand voltage examples are only true for thisconfiguration. Add another coax line or a groundedguy wire and it is completely different. (The purposeof this exercise is to show that the grounding ofthe coax at this elevated point on the tower sendsa significant amount of energy through the coaxshield towards the equipment. There is a betterway.)

THE REAL FIX!

Even though this is accepted practice, and whatyou will see most often in the field, it is incorrect.By continuing the coax further down the tower toalmost ground level and then grounding the shieldto the tower (just above the tower leg groundconnection), the instantaneous voltage gradienton the coax shield would be almost zero.Theoretically, the coax shield current would alsobe almost nothing. “Theoretically” because boththe tower ground system and the equipmentground must not only be interconnected (grounded)below grade to have this be true, but they mustalso be large enough so that ground saturationwill be minimal. Running additional ground wiresfrom the coax ground kits to the tower base willnot help either, unless you can find the theoreticalzero inductance conductor!


BEST

GOOD

OK

ACCEPTABLE

ADVANTAGES DISADVANTAGES

Low L di/dt voltage 1) Coax must maketight bends.

2) Coax enters at floorlevel.

Low L di/dt at tower 1) Large L di/dt for inline protector unlesslarge groundingsurface areaconductor is used forbuilding CGK andprotector.

2) Sloped line willintercept tower magfields.

Low L di/dt at building 1) Coax must enter atfloor level.

2) Sloped line willintercept tower magfields.

1) Enters building high. Large straps cost morebut are needed to reduce

2) Does not intercept L di/dt voltage tower mag field.

Coax Grounding Kit(CGK) In-Line

Protector

CGK

CGK In-LineProtector

In-LineProtector

CGK

CGK

CGK CGK(Built-In)

Straps

In-Line

BulkheadPanel

Protector


CHAPTER 2

Tower Strikes & Solutionsost sites in use today separate thecoax cables from the tower and routethem toward the building entrancepanel at a relatively high point on the

tower, typically 8 to 15 feet above the tower base.This practice is the single most damaging sourceof lightning energy directed toward the equipment.

Chapter 1 discussed the best way to reducecoax cable shield currents

directed toward theequipment: Take the coaxcables to the base of thetower and ground theshields there, routingthem at ground level orbelow to an entrance panel

at ground level.

TYING IT ALL TOGETHER

One grounding system must be formed inter-connecting all other site grounds. A lightning strikepossesses a given amount of current and byinstalling a radial ground system around the towerand a perimeter ground loop around the equipmentbuilding, the division of current will send most ofthe lightning strike energy to be dispersed by theradials. A below-grade perimeter loop around thebuilding will also reduce the amount of step voltageinside the loop. This is due to the repelling effectof like charge emanating from all points on theloop, reducing current flow through concretefloors, protecting both equipment and personnel.

Connecting all ground rods together forming a“single point” ground system.

The lightning pulse series is a “fast rise time” eventrequiring a ground system capable of dispersinglarge amounts of electrons into the soil very quicklywith minimum ground potential rise. This requiresmultiple paralleled inductances shunted byconductive earth. Multiple tower radials withground rods (parallel inductances reduce overallinductance, improving the transient response) havebeen proven to be the most effective method ofgrounding towers.

TOWER TO GROUNDCONNECTION

Most self-supporting or guyed tower legs areindividually bolted to plates or brackets imbeddedin a concrete pad. If the plates, brackets, or steelmesh in the concrete are interconnected (bondedtogether), a “Ufer” ground is established. Radialswith ground rods must be attached to the Ufer tofurther lower the impedance and improve theground system’s transient response.

MTower Coax Lines

Ground Rod

BulkheadPanel

EquipmentShelter

Utility Entry

Ground Rods

12 Lightning Protection & Grounding Solutions for Communication Sites

Conductors from the tower legs to the radialsystem must have low inductance (largecircumference) to direct lightning’s fast rise timepeak current towards ground with minimumvoltage drop.

The method of attachment to the legs and groundsystem should be very low resistance and notsubject to corrosion. Accepted practice is to lug alarge round conductor to a flange bolt, orexothermic weld it to the flange or bracket. Theconductor is routed, on top of the concrete pad,to where it goes below grade to the groundsystem. Emphasis has been placed on 8-inchminimum radius bends for lightning currentcarrying conductors, suspended in air, to minimizeinductance. Higher inductive right angle bends orconnections should be avoided if possible, but thealternative trade off may be worse.

If tighter bends are required to get an overall lowerinductance ground conductor, we recommendsolid bond copper strap.

Some guyed towers use a tapered base with a balljoint on a concrete pier. For example, the nearestground rod is some 24 inches (vertical + horizontal)from the top of the concrete pier. If we want toutilize the Ufer effect, connect three or four largeconductors (one per leg) from the bolts on theplate above the ball joint to the bolts on the platebelow the ball joint. Continue the same conductor(if possible) down the sides of the pier to the groundrod/radials using optimum routing as determinedbelow.

Assuming the wire gauge is #2/0, the inductanceis approximately 0.32µH/ft. Knowing the distance,the inductance for each route can be calculated.Each 90-degree bend develops about 0.1mH ofadditional inductance. Each sharp angle has about0.15µH inductance.

Solid bond copper strap is once again the betterchoice for a low inductance conductor, howeverthe interconnections to the tower base can betroublesome. Some engineers have solved thestrap interconnect problem by brazing steel tabsto leg pads (in contact with the concrete), thenturnin

exothermically welding the copper strap to thesteel tab. The tab/strap interface is coated with aweather-proof sealer to eliminate corrosionproblems. This is a good idea but, contact thetower manufacturer before drilling or welding tothe tower!

EQUIPMENT STRESS

Even with a “perfect” ground system, voltagestress during a lightning strike may still beexperienced by the equipment if the coaxial cablesare not brought to the base of the tower beforethe outside shield is grounded, or if a properbulkhead panel/ground bar connection is notutilized.

In Chapter One we examined the tower’sinductance and the associated voltage drops duringa strike. A tower connection at 15 feet above theearth may appear to be a good grounding point.An ohm meter might show it to be a “good dcground”. But it is really a poor, possibly dangerousgrounding location due to the high peak voltagespresent during the strike.

For a 1-5/8” coax cable, it is virtually impossible tomake as sharp a bend as is necessary to groundthe shield at the tower base. Yet in the absence ofother grounding methods, it is essential to groundthe shield at this point to keep the shield voltagenear zero.

Conductor leadingto buried ground

Tower Leg


If the coax slopes downward from an elevatedpoint on the tower so it enters the equipmentbuilding at or very near ground level, another shieldgrounding kit should be incorporated at the wallpenetration.

If either method is not practical and the coax mustenter the building elevated above the earth, thena large surface area bulkhead panel should be usedto ensure a low inductance route to earth ground.

BULKHEAD PANEL

Bulkhead panels have been used for many years.The initial reason was to provide the equipmentbuilding with a structurally strong entry point forone or more cables. The bulkhead panel was arigid metal plate that covered a penetrating holein the building’s wall.

In installations where the coax lines must exit highon the tower, it is best to terminate the feedercoax at a bulkhead plate/coax center pin protectorand run a smaller more flexible coax jumper fromthe protector to the equipment.

Grounding the entry bulkhead panel usuallyconsisted of connecting a ground wire to it. Inmost installations a master ground bar (MGB) isused with coax cable ground kits connected and aground wire routed from the MGB down the wallto the site ground system. A wire like this isineffective for fast rise time pulses because of itsinductance.

If a larger bulkhead plate were continuouslyextended from the entrance opening down thebuilding exterior and beneath the soil to the groundsystem, a low inductance interconnection to theground system could be made. (Because of its largesurface area - skin effect - and large W/H ratio, itshould be less inductive than an equal height onthe tower.) If the coax cable were grounded witha short low inductance connection to such abulkhead plate, lightning surge current would bestripped from the cable shield. A copper “groundingfinger” should be used and weather protected bya boot. However the ease of installation (weight)and cost of such a full length copper extensionplate may be prohibitive. A cost effective variationis to substitute copper strap material for the thickerfull length panel material going to ground, makingit lighter, easier to install, and less expensive. Thestrap would be affixed to the building with siliconeand then covered or painted for camouflage andwind resistance.

Flat strap is the best conductor for a groundingsystem. It has maximum surface area, for skineffect and low inductance. Strap actually has lessinductance than wire for a given angle bend andcan be bent to a tighter radius. Mutual inductance,

Outside/Inside Master Ground Bar(PolyPhaser - GSIE)


the cross coupling of the magnetic fields at thebend, is the reason for the added inductance of abend. The distance from one side of the strap,when bent, is further away from the opposite sideof the strap by the angle it makes, plus the widthof the strap. The distance is greater so the mutualcoupling is less. Also, the magnetic fieldsusceptibility is maximum at its edges and it issimilar to a dipole antenna. Therefore, it is lesslikely to intercept tower magnetic fields if its flatside is oriented toward the tower.

incoming coax cable and conduit circumferences.Wide copper strap will give the largestcircumference (“skin effect”) with the least amountof copper.

External cable trays or ice bridges should not be incontact with both the tower and the buildingground system. Isolate and support the cable trayat the building end. Only the coaxial cables shouldcomplete the circuit.

One can calculate expected peak current andvoltage drops for ground conductors and, if doneproperly, will get accurate inductance valuerequirements. The inductance requirements canthen be translated to physical conductordimensions. The goal is to reduce the voltage dropto a minimum. This is a valid exercise and goodengineering “practice”.

Master Ground Bar (MGB)An easier way to determine minimum grounddown-conductor sizes would be to compare thetotal of all circumferences of incoming surgebearing conductors from the tower (coax cables,conduits), to the total circumferences of allgrounding down-conductors.

If the length of the grounding down-conductor(s)from the MGB to where it goes below grade doesnot exceed the length of the coax(s) horizontal runfrom the tower, the total down conductorcircumferences should be at least the total of all

A) Copper strap may be looked at as if madefrom an infinite number of infinitely small wiresspaced infinitely close together. B) The mag fieldof each wire C) is shown and they will addvectorially D) toward the edges.

ABC

D

CENTER CONDUCTOR

Shield currents can almost be eliminated withproper grounding techniques. However, the centerconductor surge current should also be eliminatedbefore the current damages the equipment. A dcblocked type lightning protector (see Chapter 6)can prevent the center conductor’s surge energyfrom reaching the equipment if it is mounted(grounded) to the bulkhead panel. The use of a dcblocked center pin protector will prevent the sharingof differential surge energy present on the coaxcenter conductor due to high frequency roll offand velocity of propagation differences betweencoax cable shield and center conductor.

Ground conductor circumferences should equal thecombined circumferences of all coaxial cables.

(9) 7/8 coax = 24.75” circumference(3) 1/2 coax = 4.75” circumference

Total Circumference = 29.50”

2” Copper Strap= 24” circumference (surface area)= 0.9µH inductance= 2.5kV drop

Typical Entry Panel

(1) #6x8’ long ground wire= 3.4µH inductance= 8.2kV drop*

(1) #2x8’ long ground wire= 3.1µH inductance= 7.8kV drop*

* 20kA/8µs pulse


SUBPANEL

To further protect and restrict access to the coax-to-center pin protector connection, a “U” shapedsubpanel (see page 16) could be mounted/grounded to the bulkhead front plate. The subpanelis attached so it protrudes from the main panelthrough the penetrating hole inside the buildingand creates a secondary surface on which theprotectors are mounted and grounded. Allconnectors are accessible from inside the buildingfor tests and changes. If waveguide is used, itwould extend straight through, since a center pinprotector is not needed. The grounding fingerunder the weather boot (see bulkhead drawing),accomplishes proper grounding of the waveguideand coax cables. The subpanel would be deepenough for concrete block construction. The addeddepth allows for external coax feeder entry anglecorrection and jumper support in the absence ofinternal cable trays.

The bulkhead panel is made of 1/8” half hard C110(solid copper). Only this hardness of copper canbe properly tapped for screw threads. The C110copper weighs 5.81 pounds per square foot.Mounting hardware used to join the subpanel tothe bulkhead is 18-8 stainless steel.

For small to medium size sites, the bulkhead panelshould be the central grounding point inside thebuilding for single point grounding procedures.Holes can be drilled into the U panel for bondingstraps and grounding cables from inside racks ofequipment. The bulkhead panel then serves as themaster ground window or ground bus (MGB).

Other types of bulkhead entry panels are open onthe equipment side with no “U” panel for restrictedspace locations. An outside/inside ground barassembly can be retrofitted to existing sites forsingle point use. Attachment points are providedfor 6” strap external ground conductors and1-1/2” strap internal connections for groundingcoaxial protectors.

SINGLE POINT GROUNDING

Surely everyone has heard of the safety procedurethat says to keep one hand in your pocket whileworking around high voltages. If the body doesnot complete a circuit, there is no current flow anddanger is averted.

For small- to medium-size equipment rooms, it isbest to have the equipment’s input/output (I/O)protector grounds and equipment chassis tiedtogether. The telephone line protectors, coaxprotectors, and power line protectors are thengrounded either on a bulkhead panel or mountedtogether on a single point ground plate and tied tosystem ground. Equipment chassis ground wouldthen be connected by a low inductance strap tothis ground point.

An exterior ground system should consist of thetower leg grounds (radials and rods), powercompany ground rod(s), and a below-grade copperstrap sandwich bar connecting the bulkhead strapdownconductors to the below-grade buildingperimeter ground loop.

To keep equipment safe in the event of a lightningstrike, the same one-connection concept applies.Single point grounding is a grounding techniquethat ties all the equipment in a building togetherand grounds it at one common point. Implementingthis technique is quite easy.

Bulkhead/PEEP

Framing Studs on16in Centers

Typical Coax

Adjustable Built-InGrounding Strap

6in. wide x 15ft. longCopper Straps

AdjustableGrounding Finger

ConnectionsExothermic or Lug Copper Ground

Strap BarExisting PerimeterGround System

Ground Level

Ground StrapSandwich Bars

AreaHand Access

PortsEntrance


The single point ground must be implementedproperly so the coaxial protector can do its job. AsChapter One’s example illustrates, the outside coaxground plate could rise to 7.3kV above the earthground system, emphasizing the importance of asingle point grounding concept. The bulkhead or“PEEP” (PolyPhaser Earthed Entry Panel), withproper coaxial cable protectors installed, would bethe primary “firewall” protecting equipment. If theequipment has a separate path to earth ground inaddition to the bulkhead or PEEP, the added parallelpath could allow strike current to flow through theequipment.

I/O PORTS

For repeater installations with a single telephoneinterconnect, there would be three Input/Output(I/O) ports: the coax cables, power lines, andtelephone lines. These I/O’s can be either alightning source or sink. Lightning surge energymay originate from one I/O and exit another I/Ocausing circuit damage in equipment connectedto one, or perhaps all of the I/O’s. Since it isimpossible to ground an I/O, a surge protector mustbe provided for each.

The surge protector’s purpose is to divert andisolate the equipment from the surge. Whenevera surge exceeds a preset voltage, the surgeprotector diverts the surge to a ground sink. Byinstalling a surge protector at each I/O, it is possibleto configure a grounding scheme that allows theequipment to survive a lightning strike.

Bulkhead Single Point GroundPreferred method of grounding I/O protectors for coax,telephone line and power line. The bulkhead plate in turnconnects to the perimeter ground outside the equipment hut.

A single point ground system would be created, ifall the I/O surge protectors were grounded/mounted onto a bulkhead panel or MGB. Theequipment racks are also bonded to the MGB.Surge energy stripped from the lines by the coaxcable ground kits and each of the surge protectorsis diverted to ground via a single path.

Imagine each I/O port to be a hand or a foot. If ahand or a foot touched a high-voltage dc source ata single point, no current would flow through thebody and no injury would occur. (The surge currentnecessary to elevate the body up to this highervoltage might be felt.) The body must thereforebe insulated from everything else; no other pathfor current flow can exist.

As in the above example, the equipment must beproperly isolated from conductive flooring. Bymounting the surge protectors on the samebulkhead or metal plate connected to a commonground, no surge current flows between the I/O’sand no voltage drop is created (no ground loop).Damage does not occur since the equipmentchassis is also grounded to this same point. Thesurge protectors have low impedance betweenthem, so no voltage drop can develop.

PERSONNEL SAFETY & THESINGLE POINT GROUND

Why protect the equipment but risk the technician?The first and most obvious answer is he should notbe there working on equipment during athunderstorm!

A low resistance single point ground, with insulatedracks, could still allow the potential on the racks torise to dangerous levels. During a “normal” 18kAstrike with 2 or 3 return strokes of 9kA each, theevent would probably be over before the rack’scapacitance would be charged to dangerous levels.If the event were a 140 kA strike with 10 or 11return strokes of 70 kA each, the whole site wouldbe a dangerous place. It is a difficult and unpopulardecision to compromise safety no matter what thestatistics are. A technician working on equipmentduring a thunderstorm is at risk no matter whatkind of grounding scheme is used. There are some

PANEL ASSEMBLY3 PORT BULKHEAD

TELCO LINESCOPPER GROUND

STRAP

PROTECTORSCOAX

IS-PLDOPOWER

PROTECTOR

EXTERNAL PERIMETER GROUND

(below grade)INSULATE FROM FLOOR


compromises to the single point ground that canbe implemented to increase safety (at somesacrifice of system effectiveness).

• Install an overhead insulated bus bar connectedonly to the single point ground and extendingout and around the inside walls of theequipment room in a “U” shape. Do not connectany additional ground conductors downwardto the outside below grade perimeter ground.

• Connect all metallic objects within thetechnician’s reach (while touching theequipment rack) to the bus bar. The bus barground should be the only ground connectionfor the object. Even after consideringpropagation caused peak differentials, thevoltage should nearly equal the rack potential.

• Place high voltage insulating rubber mats onthe conductive floor where technicians wouldstand.

If a bus bar connected object is also connected toa different ground point, there could be additionalmagnetic field in the building as a consequence ofcurrent flow through the bus bar and connectedobject to ground. If outside low inductanceconductors (multiple copper straps) are installedat the bulkhead or MGB, the peak voltage dropwill be minimal, reducing current flow in thebuilding.

The only satisfactory approach to lightningprotection and safety is an integrated set ofgrounding techniques, protectors, and safetyprocedures all working together (You can’t engineercommon sense).

Example of Single Point Groundwithout Bulkhead

If the bulkhead plate was not installed atthe time the equipment hut was built, analternative grounding method may be used.Although the interconnect inductance ofcopper strap is greater, protectors may bemounted to a copper plate, which isconnected by strap to the perimeter ground.

Current that is diverted by the protectors shouldgo to ground by a path whose inductance is as lowas possible. If a grounded bulkhead panel with itslarge surface area, low inductance ground strapsis not used, the next best place for mounting thesurge protectors would be on an inside, floorlevel,wall-mounted plate. A low inductanceinterconnection to the perimeter ground isessential.

No matter how low the protector path inductancemay be, it still has some inductance. Since thesurge protectors send current through thisinductance, a voltage drop (L di/dt) is created. Thisvoltage drop may cause problems for sensitiveequipment. Control lines or balanced lines, forexample, may become elevated above chassisground. To ensure the equipment chassis is heldto the same potential as the surge protectors, alow inductance connection between theequipment chassis and the bulkhead panel orprotector panel is required. This conductor shouldhave a lower inductance than the coaxial shield.

PROTECTOR MOUNTING

1-1/2" Copper Strap

To PerimeterGround System

To TelcoIS-DPTLTelephone Line Protectors

Power Line

Single PointGround Copper

Plate

IS-PLDOPower Line Protector

IS-50 SeriesCoax Protector

AntennaCoax to


SHIELD CURRENT FLOW

If radials, rods, and a single point ground for siteprotectors and equipment are installed, have allthe possible problems been eliminated? Very lowsurge current could still flow on the coax shieldwithin the equipment building toward the rack eventhough it is insulated from conductive flooring.

The equipment chassis or rack, like your body, hasthe ability to accept a charge (capacitance). Therack is elevated in potential to the L di/dt (inductivevoltage drop) of the interconnection to the exteriorground via the MGB. Current must flow along thecoax jumper shield and other ground conductorsto bring the rack to this higher potential. A magneticfield is created inside the room by the lightningpulse current surge charging the rack’s capacitancethrough the coax shield and ground conductors.The path(s) from the protector panel or MGB tothe equipment radiate the field.

PolyPhaser manufactures a coax protector that dcblocks the center conductor and the shield (to 2kV).There is no dc continuity between the antennacoax shield and the equipment side shield.

LIGHTNING ELECTROMAGNETICPULSE

The fast rise time to peak current creates an electro-magnetic field (electrostatic and magnetic fields)that radiates out from the stroke discharge pathto earth. The amplitude/frequency spectra of theradiated component would depend on the currentdensity/rise time of the stroke and the distancefrom it. Frequencies in the pulse extend well intothe communications bands and can coupledamaging energy to equipment.

Amplitude spectra of the radiationcomponent of lightning discharges.

MAGNETIC SHIELDING

Lightning’s high current means that the associatedmagnetic fields from the tower will radiate and crosscouple to cable runs inside the equipment building.Sites are usually designed to have a 5 ohm groundsystem, but the building is placed close to thetower with little or no magnetic shielding. Thedistance, between the tower and the building, isusually kept small so the transmission lines areshort. This places a heavier burden on your groundsystem to absorb and quickly conduct the strikeenergy away from the tower base. Magnetic fieldsfrom the close spaced tower will cut through theequipment building causing induced damage tointerconnected equipment.

Aluminum buildings, like aluminum chassis, do littleto attenuate low frequency magnetic fields.Concrete with steel mesh or rebar, which is ferrous,will show some attenuation.

Steel shipping containers used as an equipmentbuilding, with either single or double walls will actas a faraday shield for both radiated (plane wave)RF energy and magnetic (H) fields. The containersalso provide a uniform ground for the equipmentfrom anywhere inside the container. Inside thecontainer you may not need Electro Metal Conduit(EMT) for shielding, however, for other non-ferrousenclosures you should run all electrical andsensitive lines in separate EMT conduits.

100,000

Am

pli

tud

e, µ

V/m

10Hz 100Hz 1kHz 10kHz 100kHz 1MHz 10MHz 100MHz 1GHz

Frequency

10,000

1,000

100

10

1

0.1

0.01

0.001

0.0001

Main (Return) Strokes

Step Leader


DISTANCE VS. SHIELDING

The only alternative to a ferrous container is touse distance. Magnetic fields drop off at a rate ofone over the distance (from the source) squared.To attenuate the strike’s powerful magnetic fieldfrom entering the equipment and causing upsetor damage, space the tower a practical distancefrom the building. Distance will also add length totransmission lines which gives additional seriesinductance (voltage drop), forcing more surgecurrent down to the tower ground. Moretransmission line will not pick up more magneticfield. The straight run from the tower to the buildingis orthogonal to the magnetic field from the towerand will not pick up any additional surge.

Bulkhead panel strap orientation is at minimum forH field coupling. The most coupling for both E andH fields is off the strap’s sharp edges. Therefore,the strap(s) will couple less energy than a roundconductor.

LATENT DAMAGE

Stress to electronic components can cause failureat a later date. The US military has spent large sumsof money to study what has been termed “latentdamage”. Latent damage leads to reduced MTBF(Mean Time Before Failure) of equipment.Lightning stress from coupled magnetic fields tohigh speed, small junction semiconductors, canlead to unexplainable failures. Since the user doesnot have design control over the PC board layout,trace length, proper I/O protection, or theequipment enclosure, EMP and RFI environmentalconsiderations need to be considered.

LARGE SITE GROUNDING ANDSHIELDING

At very large sites (over 30 x 50 feet) wherelightning shielding is important but steel sheetscannot be used to make a shielded room, aninternal ground halo (with multiple downconductorsconnected to non-electronic fixtures) may beprovided around the room as an inexpensive

alternative. It serves to intercept the lowfrequencies of lightning, although it is not veryeffective. It is often used for (multi-point) groundingof equipment racks. However, if also connectedto the Bulkhead or MGB, large currents would flowthrough the multiple ground conductors creatingan intense magnetic field instead of absorbing it.

GROUNDING EQUIPMENT CHASSIS

Racks are commonly used to mount larger basestation equipment and repeaters. Rack panels maybe painted or the rails where they are mountedmay become oxidized. The paint and oxidation mayhave enough resistance to prevent the rails frombeing an adequate interconnecting groundingconductor.

Under non-screen room conditions and within highRF environments, such as those found at broadcasttransmitter sites, the contact between thedissimilar metals of the bolts, rack rails, andequipment panels can form “diodes”. These“diodes” have been known to causeintermodulation interference and audio rectification.

One way to tie the equipment together is shown.Insulators support the vertical ground bus. Eachpiece of equipment is connected to the bus by ashort strap. Any “noise” created by poor joints anddissimilar metal contact within the rack is “shortedout” by the short strap. The short strap may be aresonant antenna near the frequency of the strongRF field from a nearby or co-located high-powerbroadcast. It may be resonant near your operatingfrequency, or some other intermediate frequencyused by your equipment. A grid dip meter may beused to determine whether the loop is resonantat a frequency that could cause a problem.

The loop’s resonant frequency can be found withother methods. A spectrum analyzer may be link-coupled to the loop and observed to see wherethe noise floor rises when the loop is opened andclosed. The same technique could be used with aservice monitor tuned in the AM mode to a quietchannel. (Note: Front-end overloading could givefalse readings.)


A possible alternative uses corrosion resistantconductive plating or metal treatment on the rackrails and chassis mounted rack brackets.

GROUNDED SCREEN ROOMS

Screen rooms work best for shielding equipmentfrom high RF and electromagnetic pulse (EMP)fields associated with high-power transmitters,lightning strikes, and high-altitude nucleardetonations. However, proper grounding of thescreen room is required to meet specifications.

The screen room manufacturer should be able todetail the techniques that ensure maximum screenroom effectiveness. All I/O’s to the screen roomshould be filtered and protected. All protectorsshould be mounted to the outside wall of a doublescreened enclosure.

thinginnn

Each equipment chassis in a rack is connectedby short strap to a vertical grounding bussuspended by insulators.

MULTI-POINT GROUNDING

Some microprocessor controlled equipmentmakers have attempted to reduce noise and RFon the logic bus (when connected to RFequipment) by using multi-point groundingtechniques.

When the site is a large installation, it is not alwayspractical to install a low inductance interconnectionback to the single point ground panel (unless ascreen room or container wall is used).Connections to an installed “halo loop” (an isolatedconductor run high on the inner walls connectedback to the single point ground with additionalspaced, downward ground leads to the belowgrade perimeter ground loop) are attempted toequalize potential and reduce noise. Thepropagation differentials during a lightning “event”between the coax jumpers connected directly tothe rack from the entry panel (MGB or Bulkhead)and the entire inductive path from the upwardconnected equipment “halo grounds” and theirdownward ground conductors, will cause currentflow and an L di/dt potential through the halo untilit is also equalized. Additional magnetic field willbe radiated from this current flow.

Therefore, independent rack interconnects fromthe top to the below grade perimeter ground loopwould seem to be the answer. The problems thatcan arise from not having a single point ground(lightning damage) can be worse than the possiblenoise from the RF equipment. Each site designwould need evaluation and compromise.

Electrons take time to travel from one position toanother (propagation time). For this reason, caremust be taken in designing the perimeter groundsystem and spacing the connecting points for amulti-point system.

ToBulkhead

Panel


FAST PROPAGATION TIME

The multi-point ground design concept places moreemphasis on the perimeter ground connections.Since the I/O’s are the only means by which currentmay enter the room, the lower the interconnectinductance to the perimeter ground by thenumerous parallel paths, the smaller the L di/dtvoltage present. The propagation time of theground system and the timing of the current inthe earth around the hut can cause problems witha multi-point ground that is not present with a singlepoint ground.

Look at how the surge propagates in the followingseries of drawings. Note that the majority of thesurge is diverted out and away from the building.Also note the time it takes to progress around thebuilding. This is why some currents traverse theparallel multi-point paths through the building inorder to get to unsaturated perimeter groundlocations. The race is between the speed of theperimeter ground loop versus the speed throughthe building. Normally if the ground system is good,it will be faster than the slightly more inductivepath through the building. You can wind up fightingyourself by adding more paths, making the voltagedrop less, which then makes the propagation timefaster through the building. This means morecurrent will traverse through the building withundesirable L di/dt drops and current causedmagnetic fields.

With a tower strike, and the perimeter ground looppropagation that follows, there could be a periodof time when the bulkhead panel or MGB wouldbe elevated in potential but the utility entry groundrod might be at a lower potential. During this period,current could flow through the ac power safetyground towards the neutral/ground bond at themain breaker panel. However, unless the strikecurrent and number of return strokes were high,the small safety wire conductor would “choke off”most current flow and not be a cause for concern.

While single-point grounding is in most casespreferred, large installations may make useof multi-point grounding to overcome theinductance of a ground interconnectconductor suspended in the air within theequipment hut. Careful design is required.

For large equipment rooms, using manyconnections to the perimeter ground is a viablemethod. As a result, some surge current will enterthe equipment room. However, with multiple pathsto ground, the total number of paths divides thecurrent. In this way the current could be reducedto a harmless amount (each L di/dt is small enoughso no breakdown occurs in the equipment).Ideally, each path from each piece of equipmentto the perimeter ground should be of equal length.This means that for a typical site, the downwardpaths to the perimeter ground system should beinterconnected to the inside halo about every twofeet! This is rarely done and is why the halo hasproblems! It is easier (and safer) to do a singlepoint ground system than it is to install multiplehalo to perimeter ground connections.


On a well-designed ground system, the strikeenergy spreads out initially from the building.

As it reaches and saturates the radial system,it will traverse the building perimeter.

Bulkhead Panel

COMMUNICATIONSBUILDING

Utility Entry

Distance

Coax Lines

Ground Rods

Tower

Utility Entry

Coax Lines

Ground Rods

Tower

Distance

Bulkhead Panel


Bulkhead Panel


Tower

Recommended site grounding systemabout to be hit by lightning.

Neglecting the coax currents, the strikeenergy moves outward from the tower basealong the radial line.

Bulkhead Panel


Utility Entry

Distance

Coax Lines Coax Lines

Ground Rods Ground Rods

Tower

Bulkhead Panel


Utility Entry

Distance

Coax Lines

Ground Rods

Tower


EQUIPMENT ROOM “UFER”GROUNDS

A Ufer ground on new building construction shouldnot be overlooked. Wire mesh encapsulated insidethe building floor can be used as a low impedancegrid. If the aesthetics of wires emanating fromvarious locations on the building floor are notacceptable, the wire mesh may be bonded to aninside bus ground just above floor level for easierdistribution. Two feet should be the maximumseparation between vertical interconnectionsbonding the ceiling halo and the mesh floor bus.The mesh should likewise be connected to theoutside perimeter ground at the same 2-footintervals.

CABLE TRAYS

Large installations may make use of cable trays tosupport overhead runs of coax lines and otherinterconnecting wires. Ground the cable tray tothe single point MGB or bulkhead. Connect thecable tray to the top of each rack. This preventsany arc-over from coax lines and reduces currentflow through the coax cable jumper connector.

There should be jumpers to make the cable trayone conductive piece. The tray is an excellent wayof grounding equipment racks together, since ithas a very large surface area (low inductance).

If PolyPhaser isolated shield coax protectors (IS-IEseries) are used, isolate the tray and racks from thebulkhead panel. The tray and racks will be groundedthrough the electrical ground connection back tothe below grade perimeter ground loop. Coaxialjumper cables and ground conductors should bespaced from sensitive, low level signal lines. Ideally,signal lines should be run in EMT conduit.

As it spreads, it loses energy dueto the spreading and I-R losses.

By the time it surrounds the building, theradials have spread out much of the energy.

Bulkhead Panel


Utility Entry

Distance

Coax Lines

Ground Rods

Tower

Bulkhead Panel


Tower

Bulkhead Panel


Tower

Utility Entry

Ground Rods

Coax Lines

Distance

Bulkhead Panel


Tower


Outdoor weather-proof cabinet grounding with aninsulated support structure should be consideredas follows:

• If using a pad mounted outdoor cabinet, all theenergy on the large surface area conduits and/or coaxial cables would be directed towardsthe cabinet (entering from the top or side) withresultant large currents through the cabinet tolocal earth ground. Below grade cabinet bottomentry with a low ground connection on the coaxwould reduce current flow through the cabinet(recommended). Entering conduits and/orcoaxial cable shields should be connected to alow resistance, fast transient response groundsystem through the cabinet’s internal lowinductance earth ground conductor. The usualcenter pin/shield propagation differentialvoltage would occur and could be blocked byan appropriate center pin protector.

• If the Cabinet is pole mounted, current flowfrom the coaxial cables shields (to the top ofthe cabinet) and conduit (going downward fromthe cabinet) would pass through the cabinet,duplexer housings, and connector panel PCBground plane. Duplexer internal groundconnections could sustain cumulative damageand PCB ground plane traces could bedestroyed. If antenna coaxial cables werebrought down the outside of the cabinet,looped up, and entering through a bottomconnector (preferred), the lowest inductancepath would be through the bottom panel ofthe cabinet to downward going conductors.However, large shield currents would flowthrough the connector shell of the surgebearing cable, traverse the cabinet’s bottomplate, and continue out the connector shell ofthe downward cable. An interconnecting cableor shield ground kit(s) at the bottom of thecabinet, between cables ahead of theconnectors, is recommended. Current flowthrough equipment would be minimized. Abulkhead type coaxial protector could be usedas a bottom feed through connector.

INSULATED SUPPORTSTRUCTURES

Wood or fiberglass support structures are not a goodidea. They are an insulator. The cabinet earthground, coaxial cables, and conduits on theinsulated support would be the only conductivepath for lightning energy. If a wood or fiberglasssupport must be used, the first step is to providean alternate conductive path down the pole toearth. A lightning diverter (lightning rod) on top ofthe pole (above the antenna) with a separate 6-inch copper strap as an earth ground conductor,would provide a low inductance/large surface areaconductive path to an earth ground system. The6-inch copper strap earth ground conductor shouldbe routed on the opposite side of the pole fromthe cables.

When large currents flow through any conductor,a strong magnetic field is developed around theconductor and can couple energy to other nearbyconductors. A circular conductor will usually besurrounded with a cylindrical magnetic field varyingin intensity as the current flow propagates downthe conductor The circular conductor’s cylindricalfield is indicative of its magnetic field susceptibility.

A copper strap will also develop a magnetic fieldclosely aligned with its physical shape. As currentpropagates down the strap, a magnetic fielddevelops close in to the flat portion and extendsout from the edges. The strap’s field pattern isalso indicative of its magnetic field susceptibility.

If downward circular cables were arrangedperpendicular to the flat side of the copper strap(opposite sides of wood pole), the magnetic fieldoverlap would be reduced and mutual couplingwould be minimized. The strap would conduct mostof the current to earth ground with little reverseEMF developed on the cables.


The best way to make the connection is with allgalvanized materials. This includes the groundingwire, cable and clamps. The galvanized wire isbonded to a copper conductor (just above theearth’s surface) that penetrates below grade to aradial system spreading the strike energy into theearth.

How high this bonded connection should be placedabove the soil, depends upon local snowfall or floodlevels. Snow’s electrical conductivity, although low,can still cause battery action from the copperthrough the surface water to the zinc by the solarheating of the wires. The joint should bepositioned above the usual snow or flood level.

The lead is dressed straight down from the highestto the lowest guy or with a slight tilt toward thetower at the top. After bonding to a guy wire, itshould be dressed downward from the lower sideto the next guy wire. Wire brush the members,removing all oxides, and then apply a jointcompound if a pressure clamp is to be used.

To ensure no arcing will occur through theturnbuckle, a connection from the anchor plate tothe ground rod is recommended. Interconnectleads that are suspended in the air, must bedressed so the bending radius is not sharper thaneight inches.

For guy anchors in typical soil conditions, use tworadials with ground rods. The radials need not bemuch longer than 20 feet each since there arelower currents due to the higher guy wireinductances. A chain link fence post can be usedas part of the system. Bond to the fence post andcontinue the radial to 20 feet.

GUY WIRE GROUNDING

Guyed towers are better at handling lightning surgecurrents than self-supporting towers. This is onlytrue if the anchors are grounded properly. Someof the strike current traverses the guy wires(instead of the tower) and may be safely conductedinto the guy anchor ground(s). With some of thecurrent conducted by the guy wires, less is availableto saturate the ground at the tower base.

Turnbuckles should not be a path for lightningcurrent. If the turnbuckles are provided with asafety loop of guy cable (as they should be), theloops may be damaged due to arcing where theycome into contact with the guy wires andturnbuckles.

The following diagram shows the preferred methodof grounding the guy wires - tying them togetherabove the loops and turnbuckles.

Optional Aux.Power Supplywith Zinc Block

Galvanized Steel (Zinc)

1/4” GalvanizedGuy Wire

Copper Rods

ConcreteBlock

U-Bolt Clamps

#2AWG Solid Copper

These connections should not be made withcopper wire. When it rains, natural rain water hasa pH of 5.5 to 6.0 which is acidic. Copper is onlyattacked by acids. Dripping water from the topcopper wire will carry ions that react with the lowergalvanized (zinc) guy wires. The reaction washesoff the zinc coating, allowing rust to destroy thesteel guy wire.

Lightning Protection & Grounding Solutions for Communication Sites 27Lightning Protection & Grounding Solutions for Communication Sites 27

CHAPTER 3

Grounding & Materialse have discussed the lightning event,

how sites get damaged, and how todirect this damaging energy to earth

ground. But what (in this context) is ground?Ground is the “sink” for electrons in a negativepolarity cloud to earth strike.

The problem is how to disperse the rapidly risinghigh frequency electron energy into the earth body

very quickly, with minimum localground potential rise.

AC power frequencyground system designs donot always disperse highfrequency energies (as ina lightning event)

effectively. A 1994 studypresented in the U.K. in 1997

compares the performance ofadjacent horizontal earth grids at

three different frequencies and examines theeffect of adding close-in vertical ground rods. Theauthors conclude that horizontal components withvertical ground rods close in to the point of injectionlower the resistance in a predictable way at powerline frequencies, but reduce ground potential riseby an additional 27% at 1 MHz!

GROUND THEORY

A lightning strike is a local event. If the earth werea metal sphere, a lightning discharge would createa measurable ringing gradient on the sphere.However, the earth’s resistance limits the currentand dissipates the energy so the event becomesa mere pebble in the pond scenario.

When a lightning strike is delivered to a groundrod driven into average to poor conductivity soil,the rod will be surface charged at a calculablevelocity factor. The charge will rise to a level wherethe concentration of E field lines cause the soil tobreak down at the rod point. This breakdown willmomentarily increase the surface area of the rod.As the charge is being transferred, depleting thesource, the E field at the rod tip will decrease belowwhere the arc is sustained. The charge continuesto disperse onto the surface of each grain of dirtsurrounding the rod. (A charge can exist on aninsulator, i.e., a glass rod after being rubbed with apiece of silk.) Because of the irregularity of thegranular surface, the bumpy E fields inhibit thecharge dispersal beyond a small range (sphere ofinfluence).

If a ground rod is in poor conductivity soil, we canequate this to a rod being suspended in air, so rodinductance must be taken into account. The voltagedrop, due to the inductive creation of a largemagnetic field is expressed as: E = L di/dt - where“di” is the lightning strike peak current, typically18,000 amps, and “dt” is the rise time,approximately 2µs.

There is a limitation to the length that a singleground rod can penetrate poor conductivity soilon its way to water table and better conductiveearth. Unless “shunted” by conductive earth, theseries inductance of the rod section in poorconductivity soil, “chokes off” current flow to apossibly more conductive lower section creating avoltage drop along the more inductive top section.The top section of the rod “breaks down” the soil.There will be saturation and local ground potentialrise due to this breakdown.

W


By having two rods connected in parallel, the overallinductance can be reduced. The spacing isimportant so each rod is able to “dump” intodifferent volumes of earth. This dictates that therod spacing be rather large (approximately 16 feetfor two 8-foot rods in homogeneous soil) so themutual inductive coupling between the rods wouldbe small. Connecting two rods to the same volumeof earth will cause saturation of that volume andlimit the passage of any further charge in the giventime span of the lightning strike.

Connecting a capacitance meter between two well-spaced rods will show capacitance is present inthe soil. The soil surrounding each rod is resistivelyseparated from the other. The cumulativeresistance represents a poor insulator separatingthe two electrodes and forms a leaky capacitor.

The only condition where both L and C can beeliminated is when R (earth resistance) = 0 ohms.(Here L is the total inductance of both wire, groundrod and the L of earth.)

Because a transmission line can simulate an earthground system when not in a non-linear arc mode,a Velocity Factor could be calculated once L and Care known.

VF = eLC

However, this assumes the earth is of equallyconductive makeup at all depths. If this type ofsoil were to be found, then calculating the surgeimpedance would be easy. This condition rarelyoccurs in nature.

Conventional ground testers are really impedancemeters. They operate in the 70 to 300 Hz rangeand express the measurement in Ohms(Z). Theywere originally designed for the electric utilityindustry to measure ground resistance to assurecircuit breaker operation should there be a groundfault. If the reactance of 30 feet of wire is <.004ohms at 60Hz, the same 30 feet could have aninductive (E = L di/dt) voltage drop of over 100kVfor a typical 18kA lightning strike.

With no available tester, the question remains:What is the ground system’s impedance atlightning’s range of frequencies/rise times?Although this impedance is still an unknown, designpractices with multiple parallel inductances, suchas radials and ground rods, can make a groundsystem with a faster transient response to quicklydisperse lightning’s fast rise time pulses.

GROUND RODS

Ground rods come in many sizes and lengths.Popular sizes are 1/2”, 5/8”, 3/4” and 1”. The1/2” size comes in steel with stainless cladding,galvanized or copper cladding (all-stainless steelrods are also available) and can be purchased plain(unthreaded) or sectional (threaded). The threadedsizes are 1/2” or 9/16” rolled threads. It is importantto buy all of the same type. Couplings look much

1

In the previous example, distributed resistance,inductance, and capacitance have been shown toexist. All can exist simultaneously. Theinterconnecting of these lumped elements is equalto that of a lossy transmission line or low pass filter.(Since earth R is high compared to the wire R, it isomitted.)

L L

R

C

Ground #2Ground #1

R

R

L

R R R

L

R

L

L

R R

C C C


like a brass pipe with internal threads and allowtwo rods to be joined (threaded) into each end.

Theoretically, one ground rod with a 1” diameterdriven in homogeneous 1,000-ohm per meter(ohm/meter) soil for one meter would yield 765ohms. Driving it two meters into the soil wouldgive 437 ohms. Going to three meters, however,does not give as great a change (309 ohms). Onewould get faster ohmic reduction and easierinstallation by using three rods, each one meterlong, giving 230 ohms compared to that of onerod three meters long. This assumes they arespaced “greater than the sum of their lengthsapart”. If the bare interconnecting wire is alsoburied below the surface, then the ground systemmay be less than 200 ohms. (Having one deepground rod, 40 feet or more, even if it reaches thewater table, will not act as a good dynamic groundbecause the top 5 to 10 feet will conduct most ofthe early current rise and could become saturated.Eddy currents will form in this top layer and causethe rod’s inductance to impede the flow of currentto any further depth.)

GROUND SATURATION

The statement that rods should have a separation,“greater than the sum of their lengths apart,”originates from theory, and the fact almost allground rods will saturate the soil to which theyconnect. A ground rod connects to localized,irregularly sized, three-dimensional electricalclumps. Depending on the soil make-up (layering,etc.), the volume of earth a ground rod can dumpcharge into can be generalized as the radius of acircle equal to the length of the rod at the circle’scenter. This is known as the sphere of influence ofthe rod. The sum of the driven depths of two rodsshould be, theoretically, the closest that groundrods can be placed. Anything closer will cause thesoil (clumps) connected in common to saturate evenfaster.

Correct Spacing

Incorrect Spacing

8Ft. Rod - 8Ft. Radius Influence10Ft. Rod - 10Ft. Radius Influence

GROUND RODSPHERE OF INFLUENCE

Theoretical resistance change for additionallyspaced ground rods.

0 1 2 3 4 5 6 7 8 9 10100%

70%60%50%

40%

30%

25%

20%

RE

SIS

TA

NC

E O

F M

ULT

IPLE

GR

OU

ND

S

NUMBER OF GROUND RODS

5' SPACING

10' SPACING

20' SPACING

40' SPACING

100' SPACING


Ground Rods

Interconnecting Wire (Radial)

INDUCTANCE

A sandy area has a water table at the 10-foot level.Two 10-foot ground rods are coupled and are tobe driven to a total depth of 20 feet. A second rodis to be driven no closer than 20 feet to the first,but 40 feet would be according to the “sphere ofinfluence rule. The rule can be looked at two ways:

(1) Only 10 feet of each 20-foot rod is in conductivesoil (the top 10 feet of each rod is in non-conductivedry sand), so 10 feet + 10 feet = 20 feet apart.

(2) Without taking the water table depth intoconsideration, 20 feet + 20 feet = 40 feet.

Following the rule, the separation in #2 will notwork! The interconnecting inductance will chokeoff the higher frequency components of thesurge’s rise time and create an L di/dt voltage drop.Due to the interconnecting inductance, most ofthe surge will never reach the second ground rod.Following #1’s spacing, the inductance will be less,but there are two other solutions to this real lifeproblem. First, using copper strap can reduce theinterconnecting conductor’s inductance. Second,by using chemical salts to increase soil conductivityaround the rods and along the interconnect path,the resistance is also reduced. For the bestperformance, use both solutions together with#1’s spacing.

GROUND SYSTEMINTERCONNECTIONS

As the spacing between vertical ground rods isincreased, the interconnecting wire will be able tolaunch or leak current into the surroundingconductive earth. Therefore, it can be thought ofas a horizontal ground “rod” connecting to verticalground rods. In highly conductive soil, one shouldnot be concerned about the inductance of such astraight wire because such a wire acts as a leakytransmission line with very high losses to the soilresistance. Therefore, wire size (skin effect) is oflittle importance, like that of rod diametermentioned earlier, as long as it can handle the I2 xR of the surge. For highly conductive soil, a #10-AWG (bare) is the smallest wire that should everbe used. This type of soil is not common.

In areas where soil conductivity is poor, such assandy soil, the #10 buried interconnecting wireapproximates an inductance as if suspended in air.This undesirable condition causes it to be highlyinductive, preventing strike current (which has afast rise time) from being conducted by the wire.Ground rods connected in this way are noteffectively utilized.

INTERCONNECTION MATERIALS

Solid copper strap should be used to inter-connectground rods in poor conductivity soil, solid copperstrap should be used. The strap may be as thin as0.016”. For 1-1/2” wide strap, the cross sectionalarea will equate to a #6 AWG wire. Greaterthickness gains only a little advantage because highfrequency currents (1 MHz) penetrate only to adepth of about 0.008” per surface, owing to “skineffect”. Although the strap width should be about1% of its length (e.g., 20 feet x 0.01 = 2.4” wide),1-1/2” strap is usually acceptable.

Connecting the strap to the rod may be done usinga clamp. (An exothermic weld is best but not alwaysavailable, check with your supplier for anappropriate mold or “one shot”) For 5/8” rods and1-1/2” strap, a clamp offers a way to achieve amechanically rigid, low maintenance connection.

GROUNDLEVEL

0’

20’

SAND

SAND &ROCK

CLAY

WATER

10’


The copper connection must be cleaned and acopper joint compound applied to prevent moistureingress. Copper clamps that bond straps and rodstogether in the soil and many other 1-1/2” strap tocable clamps (6AWG to 4/0) are available.

DRIVING THE SECTIONS

Driven rods will out perform rods whose holes areaugured or back-filled and not tamped down to theoriginal density. The soil compactness is betteraround driven rods giving more “connection” tothe rod. It will be necessary to purchase a “poundingcap” for hammering threaded rods or a bolt thatfits the coupling. By threading the coupling on tothe top end of the rod and threading the bolt intothe coupling, a “smash proof” hammering point isachieved, saving the rod’s threads.

What type and size of ground rod should be used?Most seem to choose the copper clad 5/8” x 8 or10 feet. The rod diameter should increase as thenumber of tandem rod sections and soil hardness/rockiness increases. The rod diameter has minimaleffect on final ground impedance.

Three individual tests (A,B,C). Each took a 1/2”Ground Rod which was used as a reference andset to 100%. The Rod size was increased anddifferent results are due to ground conductivityvariations.

GROUND ROD DEPTH

The total depth each ground rod must be driven into the soil depends on local soil conductivity. Soilresistivity varies greatly depending on the content,quality and the distribution of both the water andnatural salts in the soil. It is beneficial to reach thewater table, but it is not necessary in all cases.

In higher latitudes, single rods should be longenough to penetrate below the maximum frostdepth. In some cases, a total depth of 40 feet orless is necessary, with the average being 15 feet.Depth would also depend on the number of rodsand the distances between them.

RF, LIGHTNING, & SAFETY GROUNDS

A single driven ground rod, or one at each leg, isnever enough to ground a tower for lightning. Therods will immediately saturate and the local groundpotential will rise. There are three types of grounds,each required for different purposes:

• RF ground, such as an antenna counterpoise.A ground plane takes the place of the otherhalf of a normal vertical dipole. A good RF groundplane could be elevated above ground (tuned)and thus cannot be a good lightning ground. Ifsuch a ground plane is properly extended andplaced in the soil, it will no longer be tuned. Itcan then be used as an RF noise and lightningsink. Therefore, not all RF grounds are goodlightning grounds, but most good lightninggrounds are good RF grounds for lowfrequencies.

40

60

80

100

0.50 0.75 1.00 1.25 1.50 1.75 2.00

RE

SIS

TA

NC

E, %

ROD DIAMETER, INCHES

A

BC

2/0 AWGto 4/0 AWGto 1-1/2”Copper Strap

5/8”Ground Rodto 1-1/2”Copper Strap


• Lightning ground. This ground must be able tosink large amounts of current quickly (fasttransient response). The typical frequencyrange of lightning energy at the bottom of atower can be from dc to the low VHF range(<100MHz). The ground system must be abroadband absorptive counterpoise over thisfrequency range.

• Power return or safety ground for ground faults.This is a low frequency (60Hz) ground and maybe very inductive to lightning’s fast rise time, yetstill be usable for 60Hz.

The signal source for the three-stake fall ofpotential resistance measurement is a lowfrequency ac potential, usually around 100 Hz. Theelectrical safety ground is often not a goodlightning ground for that reason.

RADIALS

Some locations are rocky enough that only thehorizontal conductors can be placed below grade.Buried horizontal radials, like those used on verticalbroadcasting antennas, make an excellent RF andlightning ground system. Theoretically, four radialseach 20 meters (m) long, of #10 gauge wire, justburied will yield 30 ohms in 1,000-ohm/meter (ohm/m) soil. Eight radials would give about 25 ohms.Eight radials each fifty meters (163 feet each or1,300 feet total wire) on top of the ground or hardlyburied could give about 13 ohms in 1K-ohm/m soil.Theoretically, by adding 2m long rods (if possible)to this system, one on every radial (8 rods total),would calculate the system resistance below 10ohms. If the rods were spaced every 10m on eachradial (32 rods total), then the resistance would goto about 4 ohms. This is the theoretical impedanceat 100Hz for 1,000-ohm/m soil, which could besandy or rocky. A long radial run will not work aswell with a fast rise time lightning current pulse asmany shorter radials.

There is a law of diminishing returns for radials.As with sprinkler hoses, the amount of water, or inthis case lightning energy, at the end of a 75’ length

radial is so small going further is wasting time,effort, and material. It is recommended radials onlyhave a 75’ run (no shorter than 50 feet if possible)and then additional radials from the tower be usedto further reduce the surge impedance/groundresistance. The measured earth resistance of theradial system may be decreased, but like groundrods, you will need to double what you have to notquite halve the resistance value. The radial runsshould be oriented away from the equipmentbuilding as much as possible. In this way, thegreatest amount of energy is carried off from thetower and away from the equipment building.

Some have stated “if a radial is like a lossytransmission line, and the energy is not absorbedby the time it reaches the end of the radial, it willreflect back to the tower base.” This would seemto indicate there are not enough radials in poorconductivity soil since the soil becomes saturatedand will not absorb any more electrons. That isone more reason to install additional radials androds, not just longer radials.

Since the radial system is emulating a solid plate.The capacitance of this plate to true earth willdetermine the amount of charge that can betransfered. The resistance will dictate the timeconstant in which the plate will be elevated(saturated) above earth. Adding more radials withground rods will increase the surface area(capacitance) and decrease the resistance.

UFER GROUNDS

When building a new site, some radio installationsdo not take advantage of what is known as theUfer ground. This grounding technique cansignificantly reduce the overall ground systemimpedance. The Ufer technique can be used infootings, concrete building floors, towerfoundations and guy anchors.

The Ufer ground can be both a good lightningground and safety ground. Under a ground faultcondition, more total energy will be conducted toground than during a lightning strike, due to thelonger time required to clear the fault. Lightninghas a very high peak energy, but the duration is


very short. The Ufer has been proven to handleboth without failure.

Herbert G. Ufer, for whom the technique is named,worked as a consultant for the US Army duringWorld War II. The Army needed to earth groundbomb storage vaults near Flagstaff, Arizona. Sincean underground water system was not availableand there was little annual rainfall, Mr. Ufer cameup with the idea of using steel reinforcing rodsembedded in concrete foundations as a ground.After much research and many tests, it was foundthat a ground wire, no smaller than a #4AWGconductor, encased along the bottom of a concretefoundation footing, would give a low resistanceground. A 20-foot run (10 feet in each direction)typically gives a 5-ohm ground in 1000-ohm-metersoil conditions.

UFER GROUND TESTS

One of the most important tests performed wasunder actual lightning conditions. The test was tosee if the Ufer ground would turn the water insidethe concrete into steam and blow the foundationapart. Results indicated that if the Ufer wire was20’ minimum and kept approximately 3” from thebottom and sides of the concrete, no suchproblems would occur. (A Ufer ground shouldalways be used to augment the lightning groundingsystem and not be used alone. Radials, or radialswith ground rods, should be used together withthe Ufer. For those who are afraid to use the Ufer,think about this: The heating of the concrete is morelikely if the current is high or concentrated in a givenarea. This is known as current density “J”. The moresurface area there is to spread out the given current,the less the current density. Your tower’s anchorbolts are already in the concrete. If the groundsystem is poor, the current density surrounding thebolts could be high, turning any ambient moistureto steam, and could blow apart your concrete. Ifthe rebar is tied in to the “J” bolts, the area isincreased and the current density is reduced.(Corrosion will be reduced as well.)

A Ufer ground could be made by routing a solidwire (#4 AWG) in the concrete and connecting tothe steel reinforcing bar (rebar). Theoretically, theoutermost sections of the rebar structure shouldbe bonded together, not just tied. If tied, a poorconnection could cause an arc. Because arctemperatures are very high and are very localized,they could cause deterioration of the concrete(cracking and carbonizing) in that area.

Although possible, this has not been the case inpractice. The wire ties are surprisingly effectiveelectrical connections. “One might think that theties would fail under fault conditions. However, itshould be remembered that there are a largenumber of these junctions effectively in parallel,cinched tightly.” (IEEE Seminar Notes 1970.)

The use of large amounts of copper cable coiled inthe base of the tower (for a Ufer “effect”) has beenshown to cause flaking of the concrete and could,over time, also cause de-alloying of the rebar. Thiscan occur due to the concrete’s pH factor. Theuse of copper conductors, such as radials andground rods, outside the concrete, has not shownthese problems.

Using a small amount of copper wire, such a radial“pigtail” connection (short run in the concrete) willnot adversely affect the rebar during a typical 30-year tower life.

Graph of Rebar Versus Copper Wire

6

810

20

APR

30

JUN AUG OCT DEC FEB APR

"Rebar" System

Copper Wire

Gro

und

Res

ista

nce

- O

hms


CHOOSING THE LENGTH

The minimum rebar length necessary to avoidconcrete problems depends on the type ofconcrete (water content, density, resistivity, etc.).It is also dependent on how much of the buriedconcrete’s surface area is in contact with theground, ground resistivity, ground water content,size and length of bar, and probable size of lightningstrike current. The last variable is a gamble! The50% mean (occurrence) of lightning strikes is18,000 amperes; however, super strikes can occurthat approach 100,000 to 200,000 amperes.

Diameter SurgeConductor Inches Amps/Ft

Rebar .375 3400Rebar .500 4500Rebar .625 5500Rebar .750 6400Rebar 1.000 8150

The chart shows how much lightning current maybe conducted per foot of rebar for (dry mix)concrete. Take the total linear run of wire andmultiply it by the corresponding amperes per footto find out how long the ground conductor mustbe to handle a given strike current. Only theoutside perimeter rebar lengths of the cage shouldbe totaled.

Protection to at least the 60,000-ampere level isdesirable. This offers protection for 90% of alllightning strike events. A Ufer is only to be usedtogether with a radial, or radial and rod groundsystem.

HOW IT WORKS

Concrete retains moisture for 15 to 30 days aftera rain, or snow melt. It absorbs moisture quickly,yet gives up moisture very slowly. Concrete’smoisture retention, its minerals (lime and others)and inherent pH (a base, more than +7pH), meansit has a ready supply of ions to conduct current.The concrete’s large volume and great area ofcontact with the surrounding soil allows goodcharge transfer to the ground.

Sample tower base Rebar Assembly with #2/0stranded copper pigtails used to interconnectTower Ufer Ground to Equipment BuildingGround, Ground Rods, Radials, etc.

MATERIALS

Rods may be clad with copper to help prevent rust,not for better conductivity. Of course, coppercladding is a good conductor, but the steel it coversis also an excellent conductor when compared tolocal ground conductivity.

The thickness of the copper cladding is importantwhen it comes to driving the rod and when therod is placed in acidic soil. Penetrating rocky soilcan scratch off the copper and rust will occur. Rust,an iron oxide, is not conductive when dry, but it isfairly conductive when wet. In acid groundconditions, such as in an evergreen area, thecopper will be attacked. The thicker the coppercladding, the longer the rod will last. Some haveelected to tin all copper components in an attemptreduce corrosion. Actually, this is not a bad idea.Tin will protect the copper in acid soil. The tin willbe sacrificed in alkaline soil, but the copper willremain.

A swimming pool or garden soil acid/base testercan be used to determine the soil pH. Any groundsystem will need to be checked and maintainedon a regular basis to assure continuedperformance.

Exothermic Weld or Double Nut

To Radial Ground System or Equipment Building Ground Loop

J Bolts

Rebar Cage


#2/0 To Rebar Connection


ABOUT CORROSION

Corrosion is an electromechanical process thatresults in the degradation of a metal or alloy.Oxidation, pitting or crevicing, de-alloying, andhydrogen damage are but a few of the descriptionsof corrosion. Most metals today are not perfectlypure and consequently when exposed to theenvironment will begin to exhibit some of theeffects of corrosion.

Aluminum has an excellent corrosion resistancedue to a 1-nano-meter-thick barrier of oxide filmthat forms on the metal. Even if abraded, it willreform and protect the metal from any furthercorrosion. Any dulling, graying, or blackening thatmay subsequently appear is a result of pollutantaccumulation.

Normally, corrosion is limited to mild surfaceroughening by shallow pitting with no general lossof metal. An aluminum roof after 30 years onlyhad 0.076mm (0.0003 inch) average pitting depth.An electrical cable lost only 0.109mm (0.0043 inch)after 51years of service near Hartford, Connecticut.

Copper, such as the C110 recommended forBulkhead Panels, has been utilized for roofing,flashing, gutters and down spouts. It is one of themost widely used metals in atmospheric exposure.Despite the formation of the green patina, copperhas been used for centuries and has negligiblerates of corrosion in unpolluted water and air.

Joining copper to aluminum or copper to galvanized(hot dipped zinc) steel with no means of preventingmoisture from bridging the joint would result incorrosion loss over time. This is the acceleratedcorrosion (loss) of the least noble metal (anode)while protecting the more noble metal (cathode).Copper, in this example, is the more noble metalin both of these connections. See the NobleMetal Table (below) for a ranking of commonlyused metals.

Where the copper connection is with galvanizedsteel, the zinc coating will be reduced allowing thebase steel to oxidize (rust), which in turn will increasethe resistance of the connection and compromisethe integrity of the mechanical structure. Aluminumwill pit to copper leaving less surface area forcontact and the connection could become loose,noisy, and possibly arc under load.

MAGNESIUM 0.00 -0.71 -1.61 -1.93 -1.97 -2.12 -2.23 -2.24 -2.71 -3.17 -3.36 -3.87ALUMINUM 0.71 0.00 -0.90 -1.22 -1.26 -1.41 -1.52 -1.53 -2.00 -2.46 -2.65 -3.16ZINC 1.61 0.90 0.00 -0.32 -0.36 -0.51 -0.63 -0.64 -1.10 -1.56 -1.75 -2.26IRON 1.93 1.22 0.32 0.00 -0.04 -0.19 -0.30 -0.31 -0.78 -1.24 -1.43 -1.94CADMIUM 1.97 1.26 0.36 0.04 0.00 -0.15 -0.27 -0.28 -0.74 -1.20 -1.39 -1.90NICKEL 2.12 1.41 0.51 0.19 0.15 0.00 -0.11 -0.12 -0.59 -1.05 -1.24 -1.75TIN 2.23 1.52 0.63 0.30 0.27 0.11 0.00 -0.01 -0.47 -0.94 -1.12 -1.64LEAD 2.24 1.53 0.64 0.31 0.28 0.12 0.01 0.00 -0.46 -0.93 -1.11 -1.63COPPER 2.71 2.00 1.10 0.78 0.74 0.59 0.47 0.46 0.00 -0.46 -0.65 -1.16SILVER 3.17 2.46 1.56 1.24 1.20 1.05 0.94 0.93 0.46 0.00 -0.19 -0.70PALLADIUM 3.36 2.65 1.75 1.43 1.39 1.24 1.12 1.11 0.65 0.19 0.00 -0.51GOLD 3.87 3.16 2.26 1.94 1.90 1.75 1.64 1.63 1.16 0.70 0.51 0.00

Noble Metal Table: Accelerated corrosion can occur between unprotectedjoints if the algebraic difference in atomic potential is greater than + 0.3 volts.

L E S S N O B L E

LESS

NOBLE

Magne

sium

Aluminu

mZin

c

Iron

Cadmium

Nickel

Tin Lead

Coppe

r

Silver

Pallad

ium

Gold


Using a joint compound covering and preventingmoisture from bridging the metals can prevent jointcorrosion problems. The most popular compoundsuse either graphite or copper particles embeddedin a grease. As the joint pressure is increased, theembedded particles dig into the metals and form avirgin junction of low resistivity, void of air and itsmoisture. The use of a joint compound has beenadopted as the recommended means for joiningcoaxial protectors to bulkhead panels for non-climate controlled installations. Copper jointcompound is supplied for bulkhead panel groundstrap connections. This compound has been testedwith a “loose” one-square-inch copper to copperjoint, and can conduct a 25,500 ampere 8/20waveform surge with no flash over and no changein resistance (0.001 Ohms). Moving the loose jointafter the surge found no change to its resistance.

The connection of a copper wire to a galvanizedtower leg should be avoided even if jointcompound is used. The problem is the limitedsurface area contact of the round wire with the(round) tower leg. Consider using two PolyPhaserTK series stainless steel isolation clamps. The TKclamps will help increase the surface area of theconnection and provide the necessary isolationbetween the dissimilar metals. Use joint compoundon exposed applications of the TK clamps. For amore effective connection, use copper strap inplace of the wire with one TK series clamp. Otherconnectors are commercially available where thetwo dissimilar metals are already bonded together.

Silver oxide is the only oxide known to beconductive. (This is one reason why quality N-typecoax connectors are all silver with gold center pins.)Copper oxide is not conductive. The properapplication of joint compound will prevent copperoxidation.

If copper clad ground rods are used, be sure theoxide layer is removed. Tinned wire should not beused together in the ground with copper groundrods. Tin, lead, zinc and aluminum are all moreanodic than copper. They will be sacrificial anddisappear into the soil. It is recommended that allcomponents be made of the same external material(all tinned or all copper).

DOPING THE SOIL

Salts may be added to increase the conductivity ofthe soil, but it is a temporary solution that must berenewed every year to maintain the elevatedconductivity. Chemical ground rods can helpcapture the precipitation and direct it through thesalts, creating a saline solution dispersed into thesurrounding soil. It can also be fed from a timeddrip system, if domestic water is available.

Chemical additives, such as Rock Salt, CopperSulfate and/or Magnesium Sulfate, will help reducethe R (resistance) value so some dissipation canoccur. (Remember, power is I2R.) This will dampenthe ringing, transform the surge energy into heatand increase the size (volume) of the groundsystem. The latter two chemicals are less corrosivethan Rock Salt. Magnesium Sulfate will have muchless of an environmental impact than the othersalts. All salts will lower the freezing point of thesoil moisture, which is important at higherelevations. About 2 kilograms (kg) of salts will dope2 meters of a radial run for one year. About 5 kg(minimum) is necessary for each ground rod. Makesure the salts are watered in or they may be blownaway.

Encapsulation of radials in conductive gels or carbonmaterials is an alternative where little or no soilexists. Commercial products are available for thisuse. Acrylamide gel, Silicate gel, and Copperferrocyanide gel are listed here in the order ofincreasing conductivity; however, all involve toxicand/or hazardous materials. An easy alternative isto use concrete to make a Ufer ground.

ON ROCKS AND MOUNTAINS

If soil is rocky enough that radials are sometimesin air while spanning between rocks, theaccumulated inductance along the runs will chokeoff the surge currents. In this situation, numerousslightly shorter lengths of solid flat strap radialshave been effective. The copper strap’s sharp edgewill concentrate the E fields that are present dueto the existing L di/dt voltage drop and breakdownor arc onto the surface of the rock or soil.


On solid bare rock, strap arcing will help spreadout the charge onto the rock’s surface. A strike isusually an onslaught of electrons with like charge.Electrons repel and want to spread out. In doingso, they lose energy due to the resistances involved.Since little conductivity is present on dry bare rock,there will be little spreading in the time frame of astrike. If rain occurs before the event, then thesurface of the rock will be quite conductive andthe charge will spread out, losing energy in theprocess. The more it spreads, the more energy islost as the charge density is reduced. The use ofthe Ufer ground technique at the tower base andat the guy anchors will help spread the charge.

Be aware that low-frequency ringing may occursince the entire grounding scheme is being excited.Think of such a site as a giant vertical (parasitic)antenna being excited by a broadband (arc) noisegenerator (the lightning strike). The ringing willfurther stress the I/O surge protectors, such asthe power line and telephone line protectors.

A MOUNTAIN TOP “NON-RESONANT COUNTERPOISE”

By placing several copper straps in the soil or onthe rock, a counterpoise is created much like thoseused on AM broadcast tower sites. Even thoughthe mountain is an insulator, the radials charge uplike a capacitor and spread out the charge ontothe surface. The sharp edges of the strap will helpbreakdown the air and form arcs onto the surfaceof the rock.

This action will not affect the equipment and isbeneficial, since like arcing in the soil, it will reducethe elevated potential of the entire ground system.We are still dealing with a theoretical antenna (thetower) and a ground plane (the radials) which can“ring” when excited by an arc. Random lengths cutto odd multiples/divisions of the tower height arerecommended as part of the circuit to reduce thepossibility of “ringing.” Increasing the rock’s surfaceconductivity will dampen and dissipate the strikeenergy. This may be as simple as having light rainjust before the strike event.

Avoid the concept of drilling holes in a mountaintop and filling it with conductive material (a“conductive hole?”). Most mountain locations donot have fissures in the rock which allow water tocollect, making the fissures conductive. (Aconsultant in South America recommends settingan explosive charge in the bottom of such a holeto “fissure” the rock. Consider this at your ownrisk!)

Solid rock is not going to be any more conductivein a hole than on the surface. Consider whathappens in the hole after it has some electrons init? Since the electrons repel one another, few willenter the hole. Like water, the electrons will spillon to the surface of the rock and spread out. Unlikewater, the repulsion of electrons will mean fewerare needed to fill the hole and, once filled, thespreading on the surface will have an added force.The best way to disperse the electrons is by havinga radial ground system. Commercial products areavailable to encapsulate the radial in a conductiveconcrete-like substance.

A Non-resonant Counterpoise

Drilling and backfilling a single groundrod not recommended without additional

grounding components.


CHAPTER 4

Ground Impedanceefore one can design a properly sizedgrounding system for the required fall ofpotential measurement, the resistivity of

the soil must be known. The resistivity results willdetermine the conductor size, length, and numberof radials required. The measurement will alsodetermine how many rods are required, theirlength,and their spacing on each radial.

MEASURING YOURSOIL

A method for determiningmean value of soilresistivity (ρΕ) is shown.Four equally spacedelectrodes are driven to a

shallow depth; thepenetration depth (b) is kept

small in comparison to theinter-electrode spacing (a) where

(a) >20(b). A known AC current is circulatedbetween the two outer-electrodes while thepotential is measured across the inner pair. Thetester will provide an indicated resistance in Ohms(RΕ). If the electrode spacing (a) is in meters, usethe formula to convert to rho (ρΕ).

ρΕ = 2π.α .RΕ

This gives the mean value of soil resistivity (ρΕ ) inOhm-m. The electrode spacing (a) corresponds tothe depth of soil seen by the test current. Byvarying the electrode spacing, a profile of resistivityversus depth can be obtained. The results can bein Ohm-m or Ohm-cm and are “plugged in” to otherformulas determining the size and configurationof the copper electrodes in the grounding system.

B

MEASURING THE GROUNDSYSTEM

There is no substitute for an actual fall of potentialmeasurement on a ground system. Mostmeasuring techniques and instruments are similarand have similar faults. Present techniques utilizeequipment with steady state dc or (more often)low frequency AC current source waveforms.Neither comes close to simulating the dynamicsurge conditions (such as lightning) whereinductive voltage drops are developed. Problemswould be minimized if multiple parallel inductances(radials with rods) were incorporated in the designand layout of the ground system. Multiple parallelinductances lower the overall system inductance,improve the transient response of the system, andreduce ground potential rise during a lightning“event.”

Another way to obtain a profile of the soil is tomeasure a ground rod as you hammer it into thesoil. If no other ground conductors are present, inor along a 100-foot path, a fall of potential method(3 stake) measurement can be set up before a

Four stake method of measuring soil resistivity.

SMALL-SIZED ELECTRODES

A = 20B (APPROX)

A A A

EARTH

C1 P1 P2 C2

C1 P1 P2 C2

B


ground rod under test is inserted into the ground.The low frequencies used in most testers do nottake into account any inductance which may existin a ground system such as a rod penetrating asandy layer. The best way to determine theconsistency of your underground soil layers is toperform a preliminary fall of potential methodmeasurement and log the readings for each footthat a ground rod is driven. Plotting it shouldapproximate the Relative Earth Resistivity Curveshown below. Any large variation could meanwater/clay or sand/gravel. With this knowledge, abetter ground system can be designed for the RFproperties of the lightning strike.

Most sites have a grounding system, but it isusually an unknown. The ground system isconsidered an unknown because it has never beenmeasured or if it was measured, it has probablychanged over time. The soil resistivity variesthrough out the year because of seasonal moistureand temperature changes. Ground systemmaintenance must be performed to keep it inoperating condition.

Ground systems composed of copper and zinc arequickly eaten away in acidic soils; yet are stable inthe presence of alkaloids like concrete. Onlyaluminum is unaffected by acidic soils, but it isetched by alkaloids. Soil’s conductivity isdetermined by its water and salt content. The moresalts, the less water is required to reach a specificconductivity. At least 16% water content, byweight, is required for a soil to be conductive.

The three stake method, also known as theFall of Potential Method, is shown and is usedto measure the resistance of a single ground rod.This can be done on any four stake tester by tyingP

1 and C

1 together. The initial spacing between

electrodes P1, C

1 and C

2 for a simple electrode would

be approximately 100 feet, while for an entiregrounding system it could be 1,000 feet. The actualspacing may be increased or decreased dependingupon the size of the grounding system beingmeasured and the results of moving electrode P

2.

The goal is to move electrode P2 at discrete intervals

along a line between electrodes P1, C

1 and C

2 and

record/plot the voltage measurement. It isnecessary to locate the area of the curve wheremoving electrode P

2 has little or no affect on the

measured voltage, usually at 61.8% of distancebetween P

1,C

1 and C

2. Most modern instruments

convert voltage readings directly to (R = E/I) Ohms.(Impedence if an AC current source.)

D

EARTH

EARTH RESISTANCE

POTENTIAL PROBE POSITIONS

RES

ISTA

NC

E, O

HM

S

ROD 1 (P1,C1) ROD 3 (P2) ROD 2 (C2)

VOLTAGE SOURCE

ELECTRODE BEING TESTED

AMMETER

VOLTMETER

(a)

% R

ela

tive

to

first

one

foo

t se

ctio

n 100%

90%

80%

70%

60%

50%

40%

30%

58%

42%

33%28% 24%

21% 19% 17%16%

20%

10%

0%

1

Depth in feet

2 3 4 5 6 7 8 9 10

Area of LesserEarth Conductivity

Relative EarthCurve Resistivity

Area of GreaterEarth Conductivity

Gypsum is better than bentonite and can be addedto the soil. Gypsum absorbs and retains water anddoesn’t shrink/pull away from the conductor whendrying like bentonite. Adding 5% by weight, ofepsom salts will further insure moisture retentionand conductivity.


SURGE IMPEDANCE (Z)

In IEEE Transactions on Broadcasting, Volume BC-25, No.1, March 1979, it was established thatradials, together with rods, show a lower dynamicsurge impedance under real lightning conditionsthan the resistance measured at or near dc. Thisresults from a lightning induced ground saturationcausing localized arcing and creating a momentarylow impedance path between ground masses. Theeffective area or size of the grounding system isthereby briefly increased. The arcing occurs sinceany ground system, no matter how good, willmomentarily elevate above the global earthpotential. This temporary elevation may be due toa slow propagation of the surge through the earthand is measured as the velocity factor and timeconstant of the ground system. Obviously thelarger the impulse, the more arcing and the lowerthe dynamic impedance. It has been shown, thatthe lower the measured impedance using the dcor steady-state low frequency ac type instruments,the smaller the difference will be between themeasured and the real dynamic impedance.

GROUND PROPAGATION

As in any medium, a dynamic pulse, like RF, willtake time to propagate. This propagation time willcause a differential step voltage to exist in timebetween any two ground rods that are of differentradial distances from the strike. With a ground rodconnected to the base of a tower, the lightningimpulse will ideally propagate its step voltageoutwardly from this rod in ever-expanding circles,like a pebble thrown into a pond. If the equipmentbuilding has a separate ground rod and the powercompany and/or telephone company grounds areseparate still, the dynamic step voltage will causecurrents to flow to equalize these separate groundvoltages. If the coax cable is the only path linkingthe equipment chassis with the tower ground, thesurge will destroy circuitry while getting throughto the telephone and power grounds. (See singlepoint ground system.)


CHAPTER 5

Tower Top, Pole-Mounted,& High-RiseCommunications Sites

ower top protection requirements can bedivided into two categories: preamplifiersand repeater/amplifiers.

Tower top preamps are oftenused to obtain a lowerreceiver system noisetemperature (better signal-to-noise ratio) andovercome coaxial cable ormulti coupler losses. A costcomparison must be made

between a preamp andsmall coax versus a much

larger lower loss coax. The iceand wind loading factors for the tower

would also be a consideration.

To keep the number of wires to a tower top preampat a minimum, dc power is often injected onto thecoax cable center pin with the shield as the return.When a dc grounded (shunt-fed) antenna is usedand the preamp is physically located on the towerat the same height, with just a short length of coaxto the antenna, the series impedance differentialbetween the center conductor and shield isminimal. Some commercially made preamp systemsincorporate interdigital front end filters which aredc grounded and act as a front end protector.Preamp front end damage probability with theabove conditions is greatly reduced.

T PREAMP PROTECTION

Lightning damage is usually to the output circuitryof the preamp. During a strike, the tower, actingas an inductor, creates an instantaneous voltagedrop from top to bottom. Since the preamp housingis attached to the tower, it will rise to the samepotential as the tower. However, the centerconductor on the downward going coax cable fromthe preamp output has not yet been elevated totower potential. It can achieve this elevation twoways:

• Internal chassis grounds elevate and pass surgecurrents through output circuitry to the yetunelevated center conductor. Output circuitrycan be destroyed in the process.

• The coax cable shield, grounded to the towerat the top, center, and bottom, will share thesurge current with the tower and will couplesurge energy (both E and M fields) to the centerconductor.

Lightning surge currents from both sources willpropagate through the coax shield and centerconductor with different speeds and amplitudestoward the equipment. Unless application matchedprotectors are installed at the top and the bottomof the coax feeder, damage to preamp output andequipment input circuitry can occur. Coax cableseries impedance differential can be reintroducedif the equipment in the building is located somedistance from the entry port. An additional protectorwould be required at the equipment input port.


RF AND DC SEPARATE

The surge generated on the shield travels towardthe equipment building where it finds a dc injectorthat combines the dc and the RF. The surge willpenetrate through the injector to the dc powersupply, causing it to fail to the source voltage.

If the dc power supply has an SCR over voltagecrowbar or protector, the dv/dt action of the SCRcrowbar will be coupled back through the dcinjector and onto the coax cable. It forms abroadband step waveform, exciting the coax line.The line probably does not have a 50 ohmterminating impedance for these lower frequenciesat the preamp pickoff end. (The preamp impedanceis 50 ohms only in its operating bandpass.) Thisreflected waveform could reach hundreds of voltsat the preamp. The voltage amount depends onthe waveform, coax length and the preamp (and“bias T”) impedance.

Even if a dc continuity type coax protector wasinstalled with a dc turn on of 90 volts, it would beineffective. With a power supply voltage of 15 to48 volts, neither the preamps nor the power supplycould withstand the dynamic voltages necessaryto turn on this type of protector. (A 90-volt gastube won’t fire until approximately 700 volts underdynamic rise times.) Even if fired, the power supplywould feed the arc until a failure occurred.

A SOLUTION

One way to solve this problem is to make aprotector that separates the RF from the dc,protects each, and recombines the two togetherall in the same enclosure.

Even with an injector and pickoff combination, thesurge current must enter the equipment buildingand go to the rack before it can be taken back outto the perimeter ground system. A “pick-or”installed at the bulkhead or MGB decouples thedc from the rf, protects both, and recombinesthem on the center conductor. The protector issystem transparent and allows users of bulkheadpanels to prevent almost all the surge current fromever entering the building. An exception would be

an injector located at the bulkhead or MGB withdc inserted there.

The duplexing (combining) of the preamp’s outputwith multiple transmitters on one line reduces coaxcost and tower loading. At 800MHz and above,multi-channel lightning protectors for transmittersand receivers, with individual dc injectors, pick-off,and bi-directional pick-ors are available.

REPEATERS, POWERAMPLIFIERS & MICROWAVELINKS

For tower top repeaters and amplifiers, the I/O’sare the most important to protect. Telephone andcontrol lines are often overlooked. Each I/O, towertop and bottom, must be individually protected withan appropriately rated protector referenced to thelocal and single point ground. Power line protectors(ac or battery dc) must be local and single pointgrounded at the top and bottom with theequipment. The coax line protector on thepreamp’s antenna side may be eliminated if similarconditions exist as previously stated for thepreamps front end.

Above 18GHz, microwave equipment usually hasa downconverter (to an intermediate frequency)located on the back of the dish, powered throughone or two coax cables. This line(s) also handlesthe uplink and downlink frequencies as well as AFC(Auto Frequency Control) error information.Protectors are available and, like the tower toppreamps, it will take three locations, tower top,building entry, and equipment rack, to properlyprotect the system.

Whether a tower, high-rise building or water towerinstallation, the “single point ground” conceptshould be carried throughout the groundingscheme. All I/O protectors must be tied togetheron a grounded plate or equivalent near theequipment. It doesn’t matter that the equipmentwill be a few hundred thousand volts above trueearth ground. In these installations it only mattersthat a ground to the supporting structure exists,so everything will rise and fall in potential togetherwith the strike. Protectors will prevent equipment


damage by allowing survivable voltages on theI/O’s relative to the equipment’s chassis.

GROUNDING ROOF-MOUNTEDANTENNAS

Antennas on parapet walls or building tops shoulduse the building’s structural steel or existinglightning protection system downconductors. Themethods discussed below can be utilized with asingle point ground design. Elevator shafts usedto be a grounding means, but with micro-processors on board, a diverted strike could becostly!

HIGH-RISE BUILDINGS

The same single point grounding concept,previously discussed, will work for high-rise placedequipment. Tall buildings usually are steel-framedso grounding is reduced to finding a convenientlocation to ground the single point ground panelto structural steel. If the building has no structuralsteel, locate the equipment room near a verticalutility corridor. A utility corridor is a vertical shaftthat runs the height of the building. Find aconductive water pipe or route a large copper strapor 750 MCM cable to the basement. A separatedirect ground connection is made in addition tothe normal power safety ground. This will providea good ground path for the surge energy. Singlepoint grounding is the only way to protect yourequipment inside the room.

If none of the above options can be utilized,consider finding and attaching to reinforcing barin the concrete. In some countries this is anaccepted way of grounding when there is nostructural steel. With literally thousands ofinterconnections between rebar bundles incasedin concrete extending down toward the concretefooters, a good earth connection can be assured.

The single point ground is a “Main Ground Bar”(MGB) installed on a vertical structure near theequipment. All coaxial cable ground kits, cable traysand the neutral (X0) terminal of the isolationtransformer (if used) are bonded to this MGB. ThisMGB should be bonded to one or more of the

following four options. High-rise building groundingoptions, in order of preference, would be:

Structural steel will absorb a portion of lightning’sfast rise time pulse, distribute the energy throughthe steel building members, and provide a lowinductance path to the building footings and earth.The overall time the equipment would be subjectto high potentials, referenced to the outside world,would be minimal.

Concrete-encased steel reinforcing bar will absorba portion of lightning’s fast rise time pulse,distribute it throughout the structure, and providemany parallel conductive paths to the buildingfootings and earth ground. The overall time theequipment would be subject to high potentials,referenced to the outside world, could be slightlygreater than a building with structural steel.

A large steel water pipe will absorb less oflightning’s fast rise time pulse and come up to ahigher potential, referenced to the outside world,more quickly and stay there longer than either ofthe above two methods.

A single “Down-Conductor” from rooftop to anexternal ground rod(s) is unfortunately the waymany high-rise sites are connected to earth. In thiscase the grounding conductor is essentially“suspended in air” and the inductance of theconductor rises to its “free space” value. A fastrise time lightning pulse creates a rapid rise to peakpotential, referenced to the outside world, andstays there for an extended time until the down-conductor has time to drain away the charge.

The magnetic field formed around the single current-carrying conductor in free space immediatelyimpedes the flow of electrons toward earth. Sincethe magnetic field is sustained by current flow fromthe high rooftop potential, it will limit the current-carrying ability of the down-conductor until thecharge has been almost entirely drained .

The more time the rooftop equipment is exposedto a high potential, referenced to the outside world,the more possibility there is of damage. Singlepoint grounding techniques and appropriateprotectors are imperative to survival!


CHAPTER 6

Coaxial Cable LightningProtectors

ntenna manufacturers utilize shunt-fed dcgrounded antennas as a means ofimpedance matching and providing

some form of lightning protection to theircustomers. It has been proven that these antennas

do work and should be used as a meansof diverting a portion of the direct

strike energy to the tower andits ground system.Unfortunately thisprotection is designed tohelp the antenna surviveand not the equipment. Adirect hit, or even a near

hit, can “ring” an antennawhether it is grounded or not

since it is a tuned (resonant)circuit. The ringing waveform will

contain all resonances that are present in theantenna and its coax phasing lines. This meansboth “on frequency” ringing and other frequenciespresent will be propagating down the transmissionline towards the equipment. The “on frequency”energy will not be attenuated by a high Q duplexfilter or a 1/4 wave grounded stub being used as aprotector. In both instances, the “on frequency”energy will pass right through. Also, if we look at atypical dc grounded/shunt-fed antenna at the topof our 150-foot tower example, both the centerconductor and shield will be at the same 243kVpotential above ground at the antenna feed.Although the grounded antenna will help preventarc over of the transmission line, it will have a 6kApeak current traversing its length. The same paralleltower segment will have 12kA. The shared strikecurrent, between the tower and the coax, willcontain mostly low frequency components.

A The lack of high frequency components is due toboth the grounding of the antenna and theinductance of the tower/coax, which acts as a“filter.” The antenna ringing voltages, with muchhigher frequencies, will ride on top of these lowerfrequencies towards the equipment. A non-grounded antenna will arc over between centerpin and shield, creating major high frequencycomponents that will traverse the transmission lineto the equipment.

If the coax line were left unterminated as it reachesthe master ground bar, the coax could arc overthe center conductor to shield even if a groundedantenna were used. This is due to the differencein series impedance at lightning frequenciesbetween the shield and center conductor and theadditive ringing voltage. It is important to eliminateor stop this energy from being delivered to theequipment. Since coax lines are rarely left unused,(especially connected to an antenna) thesevoltages will be converted to current either by adc continuity coaxial cable arrestor, a shunt fedcavity, or by arcing over dc blocking capacitorsinside the equipment.

Contrary to popular belief, lightning energy doesnot “disappear” in the arrestor/protector box.Simply connecting a protector in series with thecoax line and expecting protection from a strike iswishful thinking. Only a properly installed andgrounded coax center pin protector can offer anymeasure of equipment input protection.


THE NEED FOR PROTECTION

Skin effect is a physical phenomenon that relatesto the limited penetration into a conductor of anRF signal, according to its frequency. This effect ispresent in coax cable, keeping the RF signal insideand any coupled outside interference on theshield’s outer surface. The effect begins to fall apartas the frequency is lowered and the penetration,which is a gradient, begins to mix the shield’soutside interference energy with the desiredinside energy. A ground loop, which imparts 60 Hzonto a desired signal, is causing ac current flowbetween ends on the coax shield due to dissimilarground potentials and is low enough in frequencyto couple energy through to the center conductor.

With lightning, the main frequency range is dc toabout 1 MHz (-3dB). This is in the range that affectsthe coax transfer impedance. The thicker the shieldmaterial, the less the effect of these low-frequencycurrents.

A test was performed on 50 feet of LMR1200(7/8”) coaxial cable typically used as a feeder. Thecenter conductor and shield on the surge side wereshorted to simulate a shunt-fed antenna. Thecurrent from the resulting voltage drop across two0.001 Ohm current viewing resistors at the far endof the cable was viewed using an HP-54522COscilloscope. The coaxial feeder assembly waspulsed with a Haefely Psurge 6.1 surge generatorwith PHV 30.2 combinational waveforem plug-inmodule. The surge generator was set for acombinational waveform output of 1.2 x 50 µSecat 6kV open circuit voltage and 8 x 20 µSec at3kAmps short circuit current (in accordance withIEC 1000-4-5 and IEEE C62.41 specifications). Thepeak output voltage and current indicated on theHaefely were 4300 volts and 1750 amps. (SeeFigure 1.) The resulting peak currents on the shieldwere 1531 Amperes positive and 688 Amperesnegative. The currents on the center conductorwere 234 Amperes positive and 63 Amperesnegative. Both the shield and center conductorreturned to pre-surge levels within 2 oscillations.A slight propagation delay was noted on the centerconductor’s peak current referenced to the shieldpeak current.

The above pulse was used on a 50’ long, 7/8”coax feeder. One end was shorted to simulate ashunt-fed antenna, while the other end went toseparate 0.001 Ohm current viewing resistors.

Figure 1

SCOPE

A

HP54522C

SHORT

GEN.

SURGE

50 FOOT COAX

TWO 0.001 CURRENT VIEWING RESISTORS

B

Ω


The same test was performed on 6 feet ofLMR600 (1/2”) coaxial cable typically used as ajumper. The jumper assembly was pulsed with thesame combinational waveshape. The Haefelyindicated peak voltage and current outputs were1020V and 2940A respectively. (See Figure 2.) Theresulting current on the coax shield was 1875Amperes positive and 563 Amperes negative. Thecurrent on the center conductor was 969 Amperespositive and 156 Amperes negative. Both the shieldand center conductor returned to pre-surge levelsafter 1 oscillation. A slight propagation delay wasnoted on the center conductor’s peak currentreferenced to the shield peak current.

A six-foot-long 1/2-inch coaxial jumper cable withthe same pulse applied as in Figure 1.

Figure 2

We should not be surprised by the above results.After all, even the manufacturer calls coax“unbalanced cable!” The current rise time at thetop of a feeder coax attached to a tower would bemuch faster, perhaps 1 or 2 µS during a lightningstrike. The differentials between shield and centerconductor with a faster pulse rise time would bemuch higher. Since lightning frequency pulses travelthrough both the different impedances of shieldand center conductor, the larger circumferenceshield will have lower inductance, therefore a faster

current rise time than the center conductor. Sincethe pulses arrive through different impedances, adifferential voltage would occur across the shieldand unterminated center conductor.

In the first example, using a 50-foot length offeeder coaxial cable, the positive peak differentialbetween the center conductor and shield currentswas 1297 Amperes, and the negative peakdifferential was 625 Amperes. If terminated to acapacitively coupled circuit (high impedance atlightning frequencies), the center conductorvoltage would quickly rise and “arc through” theequipment input back to shield potential. Ifterminated in an inductively coupled circuit (lowimpedance at lightning frequencies), current flowon the center conductor would continue throughthe inductive coupling “loop” back to shieldpotential. High peak current flow through the inputcircuit could destroy the input connector, thecoupling “loop,” and continue through to the nextstage(s). Obviously, this pulse differential must beequalized and prevented from entering theequipment!

A coax cable center pin protector could beconsidered a very fast voltage sensitive (gas tube)or frequency discriminate (filter) switch. When agiven threshold voltage is exceeded for a gas tubetype protector, the protector “switches” the energyfrom the center conductor to the shield (ground).When a filter type protector sees the lower lightningfrequencies (out of its passband), it directs themto the shield (ground). In both cases equalizationoccurs between the center conductor and theshield.


DC CONTINUITY ARRESTORSSHARE LIGHTNING SURGE WITHEQUIPMENT

Lightning arrestors with dc continuity, such as anair gap, simple gas tubes, and 1/4 wave shortedstubs, cannot divert this strike voltage differentialwithout sharing some of it with the equipment.This “sharing” for dc continuity coaxial gas tubearrestors occurs during the time period betweenzero volts and when the threshold for turn-on hasbeen achieved. Expect a short, high-voltage “spike”to occur at the output before the gas in the tubehas time to ionize and become conductive (a shortduration 700 to 1kV peak occurs with a 3kA,8/20µs waveform test pulse, and the arrestoroutput connected to a 50 ohm load. See Figure3). This high peak voltage goes to the equipmentcausing arcing, degrading capacitive inputs, orcreating damaging current flow in shunted inputs.

waveform and the stub output terminated to a 50Ohm load. See Figure 4.) This is due to its inherentL di/dt inductive voltage drop, along with perhapsmaking the on-frequency antenna ringing voltagesgreater, because of its own high Q ringing. A higherpeak voltage will be present if the equipment hasinternal capacitive coupling to the center conductorof the coax line. If it doesn’t, (e.g., a shunt-fedrepeater duplexer) the lower frequency voltagesare immediately converted to a current. In this case,dc continuity type arrestors would be relativelyuseless in stopping the surge current since thegas tube arrestor would not turn on in time andthe 1/4 wave stub would share surge current withthe equipment.

For 1/4 wave shorted stubs, from 2GHz and down,the inductance of the stub will still allowconsiderable voltage to be presented to theequipment input. (+6Vpeak, -2Vpeak ringing forthe entire test pulse waveform measured for a1900MHz 1/4 wave stub with a 3kA 8/20µs test

Non dc blocked gas tube protector. Observe 788Volt peak pulse before gas can ionize and becomeconductive. This voltage could be applied directly tothe equipment input.

DC BLOCKING IS THE ANSWER

PolyPhaser’s dc blocked filter type arrestors (seeFigure 5), when tested with the same pulse in thesame configuration as described above, will typicallylet through less than 500 milli-volts peak for lessthan 10 nanoseconds!

Quarter wave stub. Much lower peak to peak voltagethan gas tube (8 Volts peak to peak), but much longerduration. Total energy delivered to equipment inputdependent on strike event duration.

Figure 3

Figure 4


THE BEST PROTECTOR

The most effective type of lightning arrestor is “dcblocked.” There is no center conductor continuityfrom connector pin to pin. This internal capacitivecoupling prevents the sharing of low-frequencysurge current with equipment and limits thethroughput energy to an amount that can becoupled only by the electrostatic field in thecapacitor. This allows the dc blocked gas tube type“Impulse Suppressor” to fire as the voltage reachesthe turn-on threshold.

PolyPhaser has given considerable attention to thegas tube design to insure that, when transmitting,the RF power will not keep alive the gas in thetube after a strike. Many other protectors, eventhose licensed by our patent, use a type of gastube that will not extinguish properly. Thetransmitted energy continues to excite the tubewhich becomes a broadband noise generator andwill burn up unless transmit power ceases.

Some arrestors use an internal grounding coildesigned to drain any coax voltage build-up. (Therewould not be any, if a dc grounded antenna wereused.) The coil is in parallel with the gas tube anddoes not help filter higher frequency componentslike antenna ringing, etc. This type of design usesa simple gas tube and has the gas tubeextinguishing problem.

An additional problem of this design is the coil,which has added insertion losses, resonances andis wound on a ferrite torrodial core. When a hitoccurs, the coil’s magnetic field orients the domainsof the ferrite core and degrades the inductancevalue of the coil, causing further RF losses witheach successive hit. (Over 90% of the strikes areof the same polarity, so the effects of repeatedhits are cumulative to the ferrite core.) PolyPhaseruses only air core coils where they are required.The coils carry a very small inductance and createa low L di/dt voltage drop.

If a grounded antenna can’t be used and voltagedoes build up, it will not get to the equipment. Asthe protector reaches threshold for turn-on in a dcblocked circuit, it will go into a momentary soft

SURGES DAMAGE DUPLEXERSAND ISOLATORS

Not all duplexers have shunt feeds, but those thatdo can handle some of the lower frequencylightning surge current if properly grounded. Itdepends on the length of the cavity (frequencyband), the size of the shunt-fed loop and itsridgidity. (It is really best to prevent the lightningenergy from ever entering the equipment building,let alone the equipment itself.) Large magneticfields can be generated in the duplexer that canbend the loop, de-tune the cavity, and allow evenstronger magnetic fields to exist in subsequentstrikes. The strike can also weld the cavity inputconnectors together so the coax line cannot beremoved. The “on-frequency” antenna ringing cancreate large voltages inside the cavities and causeinternal arcing. If the first piece of equipment seenby the incoming low-frequency coax surge is anisolator, with each strike (if it survives) a gradualincrease in insertion loss will occur due to the surgecurrent’s magnetic field re-orienting the isolator’smagnetic field, and/or changing the magnet’s fluxdensity.

Figure 5

Dc blocked filter type protector lets through only0.318 Volts peak to peak for less than 10nanoseconds.


• Connector body plating of silver or whitebronze with a minimum plating thickness of6µ m (0.0002”).

• Avoid use of stainless, nickel, or ferrite insignal path. Use gold center pins.

• Quality machining - minimum finish of .4mm.

• Properly designed interface at connectionpanel, and contact surfaces.

• Avoid crimp connections - all connectionsshould be soldered. Clamp and solder outercontacts for best static and dynamicperformance.

• DIN connectors are less susceptible toIntermodulation than N connectors.

• Avoid hermetic seals containing Kovar.

Lightning Surge Current Ratings. Surge currentratings on coax lightning protectors are likehorsepower ratings for cars. Is more better? Somemanufacturers point to a 50kA rating and say theprotector will take 50+ strikes at 50 kA beforefailure. Although this is interesting, you might alsoask how much energy (with a 50 kA strike) doesthe protector let through to your equipment? Thestandard test for any coax protector is a 3kA 8/20microsecond waveform pulse (other standardpulses are being introduced such as a 10/350 or10/1000), with the output connector terminatedin to a 50 Ohm resistive load. The let throughenergy is calculated from the integrated peakvoltage and pulse width.

Since the purpose of any coax protector is toequalize the center conductor potential with theshield potential minimizing current flow throughthe equipment input, how much lightning currentwill actually be on the center conductor of yourcoax line? To answer this often-asked question,we need to determine two things:

• How much current is available?

turn-on as the gas barely ionizes and bleeds thestatic charge to ground. This does not create noisesince it will not get to the arc mode and lasts onlya short time.

Have it both ways! A dc blocked rf path, withisolated and protected low-voltage injector, pick-off,or pass-through ac/dc for tower top powereddevices. A series of protectors designed for receive-only from 50 MHz up and power handling transmit/receive protectors from 800 MHz up are available.This could be a “bias T” replacement that includesrf and dc protection.

Filter Type Protectors. A laser-cut spiral inductoron the surge side effectively grounds the dc andlow-frequency lightning components, while allowingthe desired frequency range to pass through a flatplate series capacitor to the equipment side (dcblocked, low “Q”, wide bandpass). Productfrequency ranges (at this writing) are from 800 MHzto 6 GHz, with ranges and bandwidths designatedby model number.

Intermodulation. Careful attention should be paidto intermodulation distortion specifications on allcoaxial products due to increased equipmentdensities and closer frequency assignments atCellular and PCS sites. Peak power requirementsare considered to assure adequate “headroom” fordigital modulation techniques.

Intermodulation problems due to non-linearity havealways been a problem. With increasing demandfor mobile communications, the need for greaterchannel capacity and more sensitive receivers hasmade Passive Intermodulation Distortion (PIM)more of a problem than ever. There are manycauses of PIM in a communications system. Onethat directly affects coaxial protector products areconnectors and connections to them. The followinglist was compiled from several articles on PIM:

• Restrict connector materials to copper andcopper alloys.


The total strike current will first be divided betweentower (lowest inductance) and all coax cables.Current on each coax will be divided between theshield and the center conductor. The shield has amuch larger surface area, therefore lessinductance, so the higher frequency componentsof the strike will easily travel on it. This means theshield will have a higher peak current with a shorterduration while the smaller and more inductivecenter conductor will have less current but longerduration.

Typically, the center conductor will have less thanhalf the total peak current. This means that whencalculated, the typical center conductor surge ona coax cable is not 40kA, not 20kA, and not even10kA. For only one coax cable and a 65kA strike(10% occurrence hit), a worst-case centerconductor peak current value would be less than7.5kA. For a cell site with nine (9) same-size coaxcables, the center conductor peak current wouldbe less than 850A each! The amount of strikecurrent on the center conductor will have a slowerrise time and lower peak current. This is importantto know since 1/4 wave stubs or other dc coupledprotectors, with a dc path on the center pin, willshare this strike current with the equipment input.

A throughput energy rating, in Joules with astandard waveshape, is a much better way ofevaluating the performance of a lightning arrestorthan knowing how many tens of thousands ofamps is required to blow it up!

HIGH FREQUENCY RINGING

Antenna and 1/4 Wave Stub. If your antenna ishit or if a strike is close to the tower, the voltagerise times at the strike attachment point can be onthe order of 20 to 50ns. This can cause antenna“ringing” since the antenna is a tuned circuit. Oncethis happens, the ringing will propagate down thecoax line on top of the low-frequency energy goingtoward the protector.

• Into how many paths will the current divideas it travels toward earth ground?

The total current available from a direct strike is agiven amount. Typically, the current will be below65kA, a 50% occurrence strike will have 18kA, andonly 10% will have more than 65kA. We will usethe 65kA figure for this discussion.

For a tower with one antenna and coax line, theamount of current delivered to the master groundbar (MGB) or bulkhead is a function of where theline leaves the tower and the length of the run tothe MGB. The higher the grounding kit is on thetower and the closer the MGB and cable entranceis to the tower, the more current will travel towardthe equipment. It is all a matter of inductance. Ifmore than one coax line is on the tower, theinductance path between the tower and theequipment will be less (inductances in paralleldivide) and therefore, the strike current to theequipment will increase. Even though the totalstrike current to all the equipment is increased,the amount on each coax line will be less (divided).

In general:

• The more coax lines there are, the more thecurrent is divided and the less there will be onany given line.

• The lower the coax is grounded to the tower,the less shield current there will be on eachcoax line.

• The lower the inductance path to ground fromthe MGB or bulkhead, the less shield currentwill enter the building.

• The farther the tower is spaced from the MGB/building entrance, the more inductance thecoax line(s) will have, and the less current willbe on the line(s). We do not recommend addingloops to increase inductance. They can couplemore energy like a transformer (depending onorientation to the tower) instead of reducingit.


A 1/4 wave stub will not reduce this on-frequencyringing and could increase the voltage since it toois a narrow band tuned circuit. In order to see thisringing, one needs a scope with a bandwidth largeenough to cover the operating frequency. Many1/4 wave stub manufacturers use 100 megasample/second scopes while looking at a 900MHz deviceand don’t show the whole picture. Observationsmade, using a 4 gigasample/second scope, with a1.1GHz bandwidth, show the ringing effects ofcellular antennas when they are hit and haveproduced the same ringing effects in 1/4 wavestubs.

The PolyPhaser filter series protectors are a low Qwide bandpass tuned circuit and not as likely toring as filters with a narrow bandpass high Q tunedcircuit.

Secondary Effects. Since the shield also has a dcpath to your equipment, the farther away thecoaxial protector is from the equipment, the morelikely it is to re-introduce the differential on thecoax line. The coupling of the shield and centerpin is what caused the differential initially. If thedistance from the protector (on your single pointground) and the input on the equipment rackexceeds approximately 20 feet, another center pinprotector should be mounted to the single pointground connection at the equipment rack.


CHAPTER 7

ac & dc Power Protectionfor Communication Sites

The incidence of damage to equipment ingeneral is higher from power line surgesthan by any other I/O port. This is not to

say more energy comes through the power line,just that the damage is more visible there. Since

the coax connection to the tower isthe source of the largest surge

current in the building, powerline port damage is usuallydue to improper groundingtechniques and lack ofsurge protection devices.There are two probableways power line caused

equipment damage occurs.

POWER LINE STRIKES

Surge current may be imposed on a power line bya lightning strike near the equipment or on anoverhead utility line. Current may directly enterburied lines when lightning strikes a street light ormay be conducted to a buried power line if lightningstrikes near the line. Whichever way the surgecurrent enters, it causes a bi-directional flow ofsurge current on the power line. Current flows bothtoward the equipment building and away from it,toward the nearest distribution transformer.

When the surge current flowing toward theequipment building reaches a distributiontransformer, part of the energy is diverted toground. Energy not diverted to ground is coupledthrough the transformer by arcing (non-catastrophic) or capacitive coupling. Surge currentcontinues toward the equipment building on boththe “neutral” and “hot” conductors of the powerline.

At the building’s main power line entrance panelthe neutral is tied to ground, reducing neutralconductor energy. Most of the energy from thelightning strike should remain on the “hotconductors”.

Since lightning is short in duration (20–350microseconds) compared to 50/60 Hz fundamentalac frequency, the surge current may either takethe form of a mostly positive ringing waveform, amostly negative ringing waveform, or a combinationof the two. It isn’t possible to state which is morelikely at any given instance because theimpedances along the path would have to bedefined. These impedances include the surgeimpedance, the load impedance and the lineimpedance.

• Surge source impedance differs on directlystruck power lines and with nearby strikes. Theimpedance changes as the surge current isconducted by arcing or capacitive coupling atone or more transformers between the strikeand the building. The surge impedance alsodepends upon the impedance of each groundconnection along the path, including thetransformers and the point at which the powerline enters the equipment building.

• Load impedance depends upon the amountof load inductance (power factor) placed acrossthe line by devices such as air conditioners,heaters and lights.

• Line impedance is governed by the length andnumber of lines, and the line resistance andtransformer impedance.

Impedance values for each of these elements areindependent variables representing an infinitenumber of possibilities.


Because the surge, load, and line impedancescannot be easily defined, the lightning-causedvoltage waveshapes are difficult to predict.Nevertheless, some standards have beendeveloped to provide guidelines for possible valuesof surge voltages and currents induced into thepowerlines by the lightning discharge.

Even though the peak values of lightning-inducedvoltages and currents are important, thewaveshape they take (i.e., rise time and duration)are critical to determine the total amount of energyin the lightning discharge.

One Standard that has been formulated isdesignated ANSI C62.41 - 1991. This standarddefines voltages/current waveforms and locationcategories for given surge exposures (Figure 1).

Demarcation between Location Categories B and C is arbitrarily taken to be atthe meter or at the mains disconnect. (ANSI/NFPA 70-1990(2), Article 230-70)for low-voltage service, or at the secondary of the service transformer if theservice is provided to the user at a higher voltage.

100 kHz Ring Wave(Figure 2)

• Category A is for “long branch” circuits suchas long run secondary ac wall outlets. Its surgewave is a 6kV open circuit voltage, and 200Ashort circuit ringing waveform (damped cosine)with 0.5µs rise time and a frequency of 100kHz(Figure 2).

• Category B is for major feeders, short branchcircuits, and receptacles located near an indoormain distribution panel. There are twowaveshapes designated to a category Blocation:

(1) 100 kHz Ringing Wave - A 6kV open circuitvoltage and 500 Amp short circuit ringing wavewith a 0.5µs risetime and a frequency of100kHz (Figure 2).

(2) Combination Wave – A wave combining a6kV, 1.2/50µs open circuit voltage, and a 3kA,8/20µs short circuit current (Figure 3 & 4).

• Category C is for outdoor overhead lines andservice entrances. The waveform is the sameas a category B combination wave, howeverthe currents and voltages are higher (6kV at3kA, 10kV at 5kA, and 20kV at 10kA). Thecurrent values are derived from the surgegenerator impedance, and the surged loadimpedance.

A 6kV 100kHz 200A & 500A ringing waveform:

Underground Services

Outbuilding

ServiceEntrance

ServiceEntrance

Underground Services

Outbuilding

TransformerMeter

Meter

Meter

ServiceEntrance

IEEE C62.41-1991 LOCTION CATEGORIES

A B C

A - Outlets and long branch circuits;All outlets > 10m from BAll outlets > 20m from C

B - Feeders and short branch circuits;Distribution panel devices.Bus and feeders industrial plants.Heavy appliance outlets with "short" connection to service entrance.Lighting systems in large buildings.

C - Outside and service entrance;Service drop from pole to building.Run between meter and panel.Overhead line to Outbuilding.Underground line to well pump.

Figure 1

0 10 20 30TIME (µ sec)

-0.8

-0.6

-0.4

-0.2

0.2

0.4

0.6

0.8

0.0

-1.0

1.0

V(t)

Vp

Dvdt = 12,000 V/µs

Rise-time = 0.5µs10µs

∴ 100k Hzdv


Combination Wave, Open-Circuit Voltage(Figure 3)

Combination Wave, Short-Circuit Current(Figure 4)

TOWER STRIKES

Surge current may also arrive to stress equipmentwithin the building with a strike to thecommunication tower. In an ideal installation, thetower, bulkhead, equipment and utility grounds areall tied together with a single point ground. Just asit is impossible to define the power line surgewaveforms accurately because many independentvariables are involved, it is also difficult to predictexactly how much stress will be delivered to thepower line circuits when lightning strikes the tower.

The tower lighting circuit is a surge producer oftenoverlooked. Protectors are available to prevent thisincoming surge energy from transferring to thebuilding’s power lines.

Approximations can be used to help predict theamount of stress expected for a single point groundsystem. If radial wires are installed with Ufergrounds at the tower base and connected to theguy anchors, surge current at the tower base isdivided by the number of radials. A short radial (onlyone) should connect the tower base to the belowgrade perimeter ground. The perimeter groundencircling the equipment building should connectto the bulkhead and not to any radials. The utilityground should connect to the perimeter ground.

With this interconnection scheme, as indicated inprevious chapters, it should be faster for the surgeto propagate via the perimeter than it is to traversethrough the building. If this isn’t the case, majorpower supply stress may occur.

Surge current coming from lightning that strikes apower line some distance from the building maybe divided several times as it flows throughtransformers and other devices, but it is importantto design protection for the “worst case.” Wherelightning strikes the low voltage secondaryconductors connected to the meter and mainbreaker, the current is divided. Current flowstoward the building and also back towards thedistribution transformer. Surge current flowingtowards the building may be in the order of 10kAor more. A power mains protector is needed toprotect the distribution panel and connectedequipment. Protectors must be able to withstandline current surges greater than those specifiedby the IEEE/ANSI standards.

The one way to theoretically limit stress on theequipment’s power supplies is to provide additionalinductance (isolation) for the power line path insidethe equipment building. Higher inside inductanceforces more surge current towards the outsideperimeter ground path. Additional inductance canbe made by placing the lines in EMT conduit, usinglong power cords to the equipment or winding coilsin the power cord. (Remember, like coax coils,

[GRAPHS FROM IEEE - ANSI C62.41 - 1991]

Duration=50µsec

0 20 40 60 80 1000.0

0.2

0.4

0.6

0.8

1.0

Time (µs)

V(t)Vp

Rise-time = 1.2µsec

0 10 20 30 40 1000.0

0.2

0.4

0.6

0.8

1.0

Time (µs)

(t)p

II

Duration= 20µsec

Front time = 8µsec


there is a limit to the voltage isolation achievablebefore breakdown.)

One possibility for damage remains in spite of thisprecuation. The phase conductor for the 60Hzpower may be at a voltage peak when the surgeoccurs, causing a breakdown. For criticalapplications, an additional power line protector isrecommended at the equipment rack.

An outlet strip type protector, that plugs into a wallsocket (the type often advertised in connectionwith computer/consumer equipment), may notprotect the equipment because it is “grounded”at the plug socket not the single point ground.The plug socket safety ground wire is “in themiddle” of two inductances (the power cordinductance from the equipment, and the safetywire inductance running back to the breaker boxand power company ground rod) and is effectivelyremoved from earth ground by the seriesinductance of the ground wire. The actual currentan “outlet strip” type protector will conduct to earthground during a strike is limited by the seriesinductance of the safety wire path and will rarelyconduct current flow close to advertised ratings.There will be a peak voltage differential betweenthe safety ground and the bulkhead single pointground during a lightning strike. The differentialcould be several kilovolts and cause damage toequipment in this current path.

For applications where there are no other groundpaths and the ac safety ground is the only groundavailable, this ground must be used. Be aware ofpossible outbound current flow to lower potentialsthrough other safety ground “protected”equipment I/Os.

PROTECTIVE COMPONENTS

There are many books now on the market whichaddress the design of components utilized in surgeprotectors. Most protectors use one or more ofthe following components:

• Air gaps• Gas tubes• Metal Oxide Varsitors (MOV)

• Silicon Avalanche Diodes (SAD)• Four-layer semiconductors• SCR’s

Air gaps handle high currents, but are slow to actand require regular inspection. Altitude,temperature, humidity, pollution, corrosion, shapeand spacing all effect air gap performance.

Gas tubes are better than air gaps, but both sharethe problem of “follow-on” current. When surgecurrent occurs during the instant that 60Hz ac isat a zero potential, follow-on current is not aproblem. If the surge arrives at any other time,triggering an arc, the arc could be sustained bythe power from the 60Hz source. The arcextinguishes when the waveform voltage fallsbelow the arcing voltage. This voltage can be evenlower if radioactive isotopes are used inside thegas tube. The US military now requires all gas tubesbe non-radioactive.

Gas Tube Arrestor

1 101 102 10 3 410

5000

200

300

400

500

600

700

dtdv

(V / µs)

ANSI C62.41, dtdv = 5000V / µs

1.2 x 50µsFiring Voltage ≈ 700V

Gas Tube ArrestorSymbol

Seal

GapInsulatingWallInert Gas(Ar, Ne)

ImpulseFiring

Voltage(V)

Cut-awayview of gastube surgearrestor(not to scale)

Electrode


A power loss of up to 1/2 cycle (8.3ms) can resultfrom a gas tube turn on that may cause problemswith sensitive equipment. The air gap and gas tubeare much like Silicon-Controlled Rectifiers (SCR);once in a conducting mode, the voltage to it mustbe almost completely removed to open the circuitagain.

There is another dilemma with air gaps and gastubes. As they “crowbar,” the dv/dt created is highin harmonic energy and can be coupled capacitivelythrough power supply transformers causingproblems in sensitive circuitry.

The metal oxide varistor (MOV) does not“crowbar,” it “clips” or “clamps,” starting at a giventurn-on voltage. MOVs handle moderate amountsof surge current and have a finite lifetime. Thedevices are made from zinc oxide granules. As thevoltage across it rises above turn-on, electronstunnel through and conduction occurs. Thegranules heat and melt together. Melted granulescannot reunite to form zinc oxide. In the end, theMOV is mostly zinc and short circuits. With heavyusage, the MOV life is shortened.

Surge currents conducted by the MOV create avoltage drop through the MOV. As currentincreases, the voltage drop rises. The action is non-linear and is often referred to as the “clampingratio.” In an ideal situation, the voltage wouldremain the same regardless of current.

High pulse current diodes (also known as siliconeavalanche diodes, SADs) come in a variety ofconfigurations. They have a more ideal clampingratio than MOVs with a faster response time tothe surge wavefront. Their lifetime is unlimited ifsurge currents remain within specified currenthandling ranges. Unfortunately, SADs do not handlemuch surge current in a single component package.Special consideration must be given to SAD surgeprotection device design to insure necessarypower handling capability. SADs and MOVs still actmuch faster than air gaps and even most gas tubes.

Gas Tube Operation Waveform

400 V/div vertical

400V

0V

-400V12.8 µsec/div horizontal

BM

I

0V

-250V

250V

5.0 µsec/div horizontal125.0 V/div verticalLINE-NEUTRAL IMPULSE

INTRATECH COMPUTER Sept 24, 1987 6:26 AM

533

V p

eak

61°

phas

e po

sitio

n3.

0 µs

ec r

ise

time

17 m

Joul

es (5

0 Ω

)

Metal Oxide Varistor (MOV)Schematic of MOV Microstructure V

I

3V

3V

MOVCircuit Symbol

Grains of conducting ZnO (ave. size "d")are separated byintergranularboundries

d

IntergranularBoundry

Electrodes

Silicon Avalnche Diode (SAD)

SAD (Cut-Away View) Bipolar SADCircuit Symbol

Zener Die(Large JunctionArea)

Nailhead Lead

Sn Pb(Solder)

OFHC Copper Lead,Solder Plated

Plastic (Thermoset)Encapsulation


The SAD’s and MOV’s speed is due partially to theirhigh capacitance. Surge current charges thecapacitance, making the effective response timeof a leadless MOV less than a nanosecond. A largepulse-handling SAD may have high capacitance. Aleadless chip SAD reacts in picoseconds.

SADs and MOVs are rarely used without leads,even though leads add inductance. Leads mustbe connected to these devices when they are usedin a protector. Protectors using either SAD or MOVcomponents are often advertised as having sub-nanosecond response times. To accomplish this,they must use lowpass filtering to offset theinductive lag, insuring a quick response. Power lineprotectors advertised as having sub-nanosecondresponse time have misleading specifications ifthey do not have lowpass filtering or leadlessdesign and installation. Filtering is important toprevent small spikes, surges and noise. Althoughthe voltage excursion of such spikes, surges andnoise may not be hazardous, it may causeequipment problems.

Another device listed is a four-layersemiconductor. It is a “follower” or “negativeresistance” device, much like the SCR. It handlesmore current than a SAD of the same size becauseof its crowbar action and unlike the SCR, it has a“turn-off” voltage.

Four-layer semiconductor protection devices arenot limited to power line applications. They mayalso be used on telephone or control lines, aloneor in complex combinations (hybrids).

The last device on the list which could be appliedin a protection circuit is the SCR (Silicon ControlledRectifier). This device comes in a variety of sizesand could be very fast when teamed up with anSAD or MOV. The SAD/MOV provides the speedand the SCR protects the SAD/MOV from longduration surges. In power main applications, theSCR’s dv/dt problems could be buffered by thecapacitance of the SAD/MOV and any filtering thatis present.

AC MAIN POWER PROTECTORTYPES

Parallel or Shunt Type Protector. The shunt typeac protector is the most common protector circuitin use today. It can be a simple MOV from eachlead to the safety ground, matched MOVs or SADsin a differential and/or common mode, or a hybridcircuit of cascading components for fast turn-onand high surge current capability. The shunt typeprotector is not load dependent and must belocated close to the entrance panel to reducepropagation delays. A 10kA rated protector in atypical installation will rarely conduct anything closeto its current rating except during a direct strike tothe secondary drop at the entrance panel. Theamount of current the protector will conduct toground depends on the equipment loadimpedance, the inductance of the groundconductor on its way to earth ground, and the fallof potential resistance/impedance of the earthground system.

Series In-Line Type Protector. An In-Line ac powerprotector consists of a pair of protective devicesor hybrid circuits per phase, connectedline toground, one on each end of a load bearing airwound (air will not saturate) series inductor. Thereshould be a circuit breaker or fuse in series witheach non-linear device or hybrid circuit that opensin the event of a failure of the suppressor stage(they usually fail shorted), allowing power tocontinue to flow to the load. If a protective devicefails, the associated circuit breaker opens and couldactivate an alarm circuit. There could be fixedcapacitors in parallel with the intrinsic capacitanceof the protective device(s) forming, with the seriesinductor, a Pi Network low pass filter. The additionalcapacitance absorbs some of the fast rise timeenergy as the device is turning on, and the lowpass filtering provides EMI/RFI filtering to theprotected lines. Turn-on time would not be an issuedue to the “filtering action” of the circuit.

If there were two MOVs (for example) in eachphase, separated by a series air wound inductor,the second MOV would be there to make a(nominal) current flow in the inductor creating a


voltage drop. This drop limits the current to theequipment side MOV and allows the surge sideMOV to take the brunt of the energy. The voltageacross the surge side MOV will rise as increasingsurge current goes through it (clamping ratio). Thevoltage difference between the surge MOV andthe equipment MOV is the voltage drop throughthe coil (E = Ldi/dt). Since the equipment MOV isnot seeing as much voltage, its current will besmaller. This is desireable since the voltage acrossthe equipment MOV will be the same voltageapplied to the protected equipment input. The lifeof the equipment MOV will be longer than thesurge MOV. The same surge current limitationscaused by inductive ground connections and/orground system resistance/ impedance apply.

A shunt type or series type protector can conductsurge energy to their capacity only if attached tothe site single point ground with a low inductanceconnection to a low resistance/impedance, fasttransient response ground system.

Applying a single point ground protector retrofitto a site with coax entry on one wall and acpower on the opposite wall?

The ground connected to the ac power protectorshould be referenced to the single point ground.Assuming a tower strike, the inductive peak voltagedrop across the current carrying ground leads(preferably copper strap) from the ground bar orbulkhead to earth would elevate the potential atthe ground bar compared to the (lower) earthground connection. The ground bar is directlyconnected to the equipment cabinet(s) throughthe coax cable shield(s).

All equipment ground connections should rise andfall in potential at the same time with no other paths(through equipment?) to a lower ground potentialconnecting point. If there are no other paths, andthe potentials rise and fall together, there will beno current flow through the equipment.

If the protector and ground connection is on theopposite wall connected to a ground rod or ring,there could be damage from current flow from theelevated potential on the coax cable shields,

through the equipment, back to the ac powerground on the opposite wall (that has not yet beenelevated in potential) due to ground potentialpropagation delay around the ground ring. Therecould be an additional danger from the energycoupled to parallel or nearby conductors. The peakpotentials could be additive and cause seriousdamage.

A protector mounted on the opposite wall next tothe main ac entrance panel, and connected to thepower company ground rod, would only protectfrom incoming energy on the secondary ac powerconductors. Energy from a tower strike wouldelevate the equipment cabinet potential via thecoax cable shields and current could flow fromelevated rack/chassis ground returns up throughpower supply circuitry on its way to the outsideworld via the yet to be elevated ac secondaryconnection. The power supply could be destroyed.

If a protector ground was connected to the(elevated) single point ground bar and true singlepoint connections were maintained:

• The equipment protector reference groundwould rise and fall with the master ground barpotential.

• The ac power protector would protect theequipment power supplies from incomingenergy on the ac power lines AND from director induced energy incoming from the coaxcable(s) during a tower strike.

One practical way to do a retrofit could be:

• Remove any equipment rack or activeelectronic equipment circuit breakers from theexisting distribution panel. Remove all wiringfrom the existing distribution panel to anyequipment rack or active electronic equipment.Leave the existing shunt type ac protectors attheir original location.

• Run a steel conduit, grounded at both ends,from the existing main distribution panel to anew sub-panel. Route new interconnectionwiring through this conduit.


• Install a new sub panel next to the single pointground (MGB). Wire from the new sub panel,through a properly rated circuit breaker, directlyto each rack or active electronic equipment.

• Connect a second shunt type ac protector tothe hot leads and neutral in the new sub panel.Strap the protector case ground to the singlepoint ground. For higher levels of protectionand RF-EMI filtering, use an in line filtered acprotector. This device installs in series with thewiring from the existing distribution panel tothe new sub panel.

If there is already ac power shunt type protectioninstalled at the main ac input, leave it there. It willprovide an extra measure of protection from anincoming strike on the secondary conductors.

AC LINE REGULATION

An ac line protector is not intended for powerconditioning or over-voltage conditions. An ac lineprotector is specifically designed to protectequipment from short duration (nano/micro-seconds) high energy “spikes” caused by lightningevents or other short term transient artifacts onthe ac line. Extended surges and sags are handledwith uninterruptable power supplies (UPS), and/orregulation transformers and/or brute filtering.Frequently, improved service grounding (earthing)can significantly improve “dirty” utility companypower.

COMMUNICATIONS SITEGENERATORS

All inside or outside generators must be connectedto the perimeter ground. Both the neutral and themetal housing are to be grounded. All fuel tanksmust be grounded, even if they are buried or tarcoated (insulated). The location of the ac mainsprotector may be changed depending on thelocation of the generator and transfer switch.

Some generators hunt/vary their frequency oroutput voltage causing equipment problems.Protectors placed on the output or load side of

the transfer switch will not remove long durationvoltage peaks. They may be damaged or destroyed.Lightning protectors should be placed on the utilityinput side of the transfer switch.

BATTERY AND CHARGERPROTECTION

Some installations use batteries in combinationwith a charger to supply power to the equipment.The charger needs power line protection to survivea surge from a lightning strike.

Batteries that are in good condition providesubstantial line-to-line capacitance, but they don’tprotect from common mode surges (lines toground). If the batteries are located near thecharger and the dc power lines to the equipmentare long (inductive), a dc over-voltage protectormay be needed at the equipment.

In non-screen room installations, long dc powerlines will pick up the electromagnetic pulse of anearby lightning strike. A capacitor bypass networkcan be used to shunt the pulse to ground. Thenetwork should have four parallel capacitorsconnected with very short leads. Values of 0.01µF,0.1µF, 1µF, and 10µF are recommended. A highpulse current SAD may also be incorporated toclamp over voltages and reverse spikes from theequipment’s dc power line. Be sure the turn-onvoltage is high enough so the battery surfacecharge will not turn on the SAD. The SAD, like theMOV lightning protector, should not be used as ashunt type voltage regulator. EMP pickup can alsobe reduced by enclosing the dc lines in metalconduit. The conduit is grounded only at theequipment end.

Batteries should not use the earth ground bus inthe equipment room as a means of providing areturn circuit. Separate conductors for positive andnegative connections should be run to each pieceof equipment or equipment noise could be spreadthroughout the ground system.


Some manufacturers ground the negative (orpositive) side of dc operated equipment to thechassis/rack. This reinforces the need for local dcover voltage protection since the chassis groundpotential follows the rest of the ground system’spotential during a lightning strike and could go up.

Stress could occur at the battery charger outputbecause of the difference (inductive delay) in thepositive and negative lead lengths and chassisground. The problem can be solved by using andMOV/SAD devices at the charger output. Thedevice’s capacitance should be balanced (equal)to prevent hum pickup. They should be placedbetween each floating output and chassis ground.

The SAD/MOV device prevents an arc breakdownfrom occurring within the charger. Such abreakdown could damage components. Choosean SAD/MOV device rating high enough to allowthe charger’s maximum dc voltage to pass.Additional components could be installed at theequipment end to provide local protection.

Connections inside the equipment building shouldbe free of paint and clean of all contaminants.Copper based joint compound is recommended.The connections should be tested when made asa preventive maintenance procedure.

At battery-operated sites, the ground connectionmay have dc current flowing through the joint undertest and a negative resistance value may bedisplayed. Whether displayed or not, the DVMwill need to measure the joint in both directions(reverse leads). The algebraic sum or the differencebetween the two readings, if both are the samepolarity, will be the real ohmic value.

SOLAR PANELS

At installations using solar-power photoelectriccells, the lines connecting them to the regulatorcan act as an antenna capturing the radiated surgefields of a lightning strike. Spike voltage line-to-line may be low because of the solar cell’simpedance, but the surge voltage from each line-to-ground may be quite high.

The high line-to-ground voltage could stress orcause premature failure of the series passregulator. Any regulator failure means eventualoutage for the equipment. SAD/MOV surgesuppressor devices should be applied to limittransient voltages.

Never attach a lightning rod to a panel support, nomatter how exposed a solar panel is to a directlightning strike. The most effective way to protectsolar panels is to use a lightning strike divertorplaced a short distance away from the panel.


CHAPTER 8

Telephone Network &Computer Interfaces atCommunications Sites

n entrance protector is required fortelephone or control lines for the samereason an entrance protector is needed

for power lines. The local telephonecompany usually supplies the

building entrance stationprotector as part of theirinstallation service. This isusually a single gas tubeper wire type protector.Ground it to your

perimeter ground system.The telco supplied

protectors prevent surgecurrents from entering the building

and radiating a magnetic field inside the equipmentroom.

If a better protector is desired, it should be used inaddition to the telco-supplied protector andinstalled after the demarcation block. Thetelephone lines should be run to the bulkhead panelin EMT conduit to prevent surge magnetic fieldpickup. Install the additional secondary protectorat the bulkhead panel for single point grounding.

The telephone lines running between the entranceprotector and the single point ground secondaryprotector are inductive and will impede fast risetime current. If the distance is short, inductancecan be increased by using metal conduit orshielded loops. The line may also be enclosed bya metal conduit grounded on the equipment endor looped several times inside a steel (ferrousfaraday shield) NEMA enclosure.

A FEWER LINES NEED BETTERPROTECTION

Dedicated lines for dc remotes, tone remotes, datalines and audio lines should be treated in a mannersimilar to the telephone line. Since the total surgeenergy is a given amount, the more line pairs thatenter the building, the less surge current per pair.A single pair, terminating at the equipment, needsa better protector (multi-stage) than each pair of a25-pair line terminating at the same equipment.

For multi-pair installation, extensive use of balancedMOV/SADs provides both protection andcapacitive filtering. The MOV/SADs should beplaced line-to-line and line-to-ground. To preventpair imbalance that could induce a differentialvoltage, both of the line-to-ground MOV/SADsshould be selected so their turn on voltagematches to within 5%. The turn on of the line-to-line MOV/SAD is not critical.

PROTECTOR PLACEMENT

Protectors should be placed on the bulkhead panelin buildings using the single point groundingtechnique. This allows each I/O protector (coax,power and telephone lines ) to have a lowinductance interconnection to the perimeterground, as well as to each other. In an installationwithout a bulkhead, the I/O protectors should beplaced on a single point grounding plate connectedto the perimeter ground.


THREE-ELEMENT GAS TUBE

The telephone company will usually supply a gastube protector at the building entrance. The gastube was originally designed to replace the oldercarbon buttons that became noisy with use andrequired maintenance. Actually, the carbon buttonsweren’t intended to provide lightning protection,but rather to protect personnel in the event a powerline came in contact with a telephone line.

Both the carbon button and the gas tube areinadequate for protecting equipment. On balancedtelephone lines both the voltage above ground(common mode) and the voltage between lines(differential mode) are important. When individualcrowbar devices are used, one for each side of asingle pair (line), usually one device fires beforethe other, creating a large differential voltage thatcan damage equipment.

One solution is to use a three-element gas tube.This device has a common gas chamber with twogaps, one on either side of the grounded electrode.When one side of the line reaches the ionizationpotential of the gas, both sides fire simultaneouslyto ground. Once again, because of the rapidchange (dv/dt) caused by the device firing, it isimportant to use some lowpass filtering.

Telephone lines, data, or control lines are similarto power lines since they provide two directionsfor the surge energy to flow. Since telephone, data,and control lines use wires smaller than ac powerwiring, they have more impedance (inductance andresistance). With higher impedance the physicaldamage caused by the surge current decreases.

VOICE/DATA I/O

Fiber optics with non-metallic armor or strengthmembers are the way to go, if it is affordable. Thiswill eliminate one I/O hazard.

If wire pairs are used, protectors are necessary.Gas tube type arrestors installed by the phone

company, will soon be outdated. The large dv/dtcreated with the crowbar action of the gas tubewill cause on-line transient problems with digitalequipment. Lowpass filtering is important to limitthe harmonic energy created by the crowbar actionof the gas tube on wire lines. Non-crowbarringprotectors such as capacitively balanced MOV/SADs are recommended.

For L-carrier coax, a coax grounding kit should beused prior to entering the building. A coaxialprotector should be used to protect the equipmentfrom the center conductor’s differential surgeenergy.

For T-carrier on pairs, (and LAN) special protectorswith high bandwidths and low (logic level) turn-onlevels should be used. Special units for telco (spanline) repeaters (current loop) are available.

EQUIPMENT ROOMPRECAUTIONS

Always use anti-static floor material to preventElectro-Static Discharge (ESD) buildup and to allowbleedoff of existing charge on personnel andmoving carts. Anti-static flooring will prevent ESDbuildup, but is of too high impedence to bleedoffpreviously aquired charge. Ground the metal poststhat support computer room raised floors. EMPgasket material should bond the cast aluminumtile squares to the structural posts. This techniquewill not only aid in ESD, but will also form a groundplane which will be beneficial for EMP attenuation,if the room is not a screen room enclosure. Othermeans such as ion generation may also be helpful.

Another problem is not maintaining a lowinductance interconnect path between protectors(L & T-carrier, power and other incoming pairs) andbetween protectors and system ground.

The most common problem found at sites andcomputer rooms is the practice of running linesthat can carry surge current together with otherlines that should be quiet. For example, groundconductors from protectors being run together orcrossing with other ground lines, battery return


lines, and data lines. The coupling of the surge tothe other clean lines can cause major equipmentglitches or even down time. EMT conduitgrounded on one end is the best way to providefaraday shielding.

COAXIAL CABLE DATA LINES

Communication lines that travel from one floor toanother, or from building to building, can pose asurge threat to equipment I/Os, as well as hum orground-loop problems.

Coaxial cable has more problems with ground-loopcurrents than a balanced twisted pair. The coaxshield is normally grounded to the signal groundat each end, which in turn, is grounded to the localpower line neutral and earth ground. Two different

IS-PLDO-120-15A

IS-PLDO-120-15A

IS-50NB

IS-50NB

LAN COAX

earth ground locations can cause ground loopcurrents to flow on the shield if there are groundpotential differences. Grounding the coax shieldonly at one end sometimes solves the ground loop,but can create other problems.

Another solution uses a PolyPhaser protector withshield continuity mounted on a ground isolator.The ground isolator can withstand up to 90V shieldto ground potential without turning on. Theincoming coax cable shield is not connected tothe remote equipment local earth ground unless a90V potential is exceeded. Protection is providedwhile still ensuring the quality of the data.

For protection at the terminal end, a standard coaxprotector with no isolator may be used (see groundloop discussion on page 73, “Security Cameras”).

IS-PLDO-120-15A

IS-50NB

LAN COAX

IS-50NB

IS-PLDO-120-15A

Lightning Protection & Grounding Solutions for Communication Sites 69Lightning Protection & Grounding Solutions for Communication Sites 69Lightning Protection & Grounding Solutions for Communication Sites 69

CHAPTER 9

Protecting Equipment fromNEMP Damage

The sudden release of gamma rays (highenergy rays) in a nuclear explosion willcause almost instant ionization (the removal

of electrons from atoms) of the atmospheric gasesthat surround the detonation. Free electrons

are driven outward. Gamma rayscan travel great distances

ionizing the atmosphere.This forced movement ofelectrons, which will againrecombine withatmosphere atoms(Compton Recoil Effect),

creats a pulsed electro-magnetic field (EMP), or

“Electromagnetic Pulse.” This isalso referred to as “Nuclear

Electromagnetic Pulse” (NEMP). About 99% of theNEMP is radiated in a broad spectrum between10kHz. and 100MHz. Most of the energy is atfrequencies below 10MHz. For comparison,lightning’s power density spectrum is from dc to 1MHz (for the -3dB point).

The fast rise time of the radiated pulse (10nanoseconds,) as well as its short duration (1.0microsecond,) can cause any antenna to ring muchlike a direct lightning strike. The ringing amplitudewould depend on the amount of captured energy.The antenna’s capture area, its pattern relative tothe blast, tuned frequency, and bandwidth, allaffect the peak ringing voltage that will be present.This ringing voltage will attempt to propagate downthe transmission line (open-balanced or coax-unbalanced) to the equipment.

Since the antenna impedence is not equal to theline’s characteristic impedance over the entireNEMP spectrum, and the line may also collectseveral

Normalized NEMP spectrum shows most RF energyfrom pulse falls at frequencies below the 10MHzamount. 99% of energy is within the spectrum from10kHz to 100MHz

Amplitude spectra of the radiation component oflightning discharges.

100,000

Am

pli

tud

e, µ

V/m

10Hz 100Hz 1kHz 10kHz 100kHz 1MHz 10MHz 100MHz 1GHz

Frequency

10,000

1,000

100

10

1

0.1

0.01

0.001

0.0001

Main (Return) Strokes

Step Leader

dB

0

-10

-20

-30

-40

-50

-60

-70

-80

-90

-100

101 210 310 410 510 610 710 810 910 1010

FREQUENCY (Hz)


NEMP energy. If additional shielding precautionsare not taken, the energy could take the form of acomplex waveform on the transmission line. Largevoltages may be created due to line and antennaresonances. These high voltages can causedamage to unprotected equipment or cause arcingin the line or at the antenna. Cable shield groundingkits help prevent the lower-frequency componentsfrom being present on coax cable and may changethe high-voltage resonance locations.

Cavity filters could increase NEMP damage. Thesmall bandwidth (high Q) of the cavity causes largerringing voltages to be present at the equipmentthan the equipment would receive if the cavityfilters were not in-line. Quarter-wave shorting stubtype protectors, depending on their “Q”, may alsoworsen the effects.

Coaxial lightning arrestors with a 50ns responsetime are too slow for the NEMP ringing pulse forsystems above 20MHz.

NEMP PROTECTORS

Only NEMP coaxial units with a 1.0 to 7.0nsresponse should be considered for protection.NEMP protectors with dc continuity must have atleast 10 feet of coax between the protector andthe equipment. Since coax has a velocity ofpropagation factor, the NEMP rise time is delayedin reaching the equipment. This gives the protectortime to operate. If this type of NEMP protector isplaced directly on the equipment, a receiver (forexample) would be required to develop an L di/dtvoltage drop across its internal static drain inductorlarge enough to allow the protector to operate.The inductor value (which depends on the receiverfrequency), the maximum current/voltage that theinductor can withstand, and the threshold ofoperation for the protector, will all determine thecondition of the receiver after the NEMP event.

NEMP protectors that do not have dc continuity(dc blocked) will work on all equipment and do notneed special precautions. They will work forlightning as well! Lightning radiates much likeNEMP, but is more localized.

Typical EMP energy collectors and responses.

Rise Time Peak Voltage Peak CurrentType of Conductor (Sec) (Volts) (Amps)

Long Unshielded Wires 10-8 - 10-7 105 - 5 x 106 103 - 104

(power lines, large antennas)

Unshielded Telephone and 10-8 - 10-6 100 - 10-4 1 - 100AC Power Line at Wall Plug 10-7 - 10-5 103 - 5 x 104 10 - 100

HF Antennas 10-8 - 10-7 104 - 106 500 - 104

VHF Antennas 10-9 - 10-8 103 - 105 100 - 103

UHF Antennas 10-9 - 10-8 100 - 104 10 - 100

Shielded Cable 10-6 - 10-4 1 - 100 0.1 - 50

POWER LINES

Utility lines act as a bandpass filter to the NEMPpulse. Transformers attenuate energy below 1MHzand capacitively couple energy to about 10MHz.Secondary ac power line protection with responsetimes of 50ns and less should be used at or on theequipment.

TELEPHONE/CONTROL LINES

Twisted pair lines can couple some NEMP energy,but not as much as the coax or power lines. Thehigher inductance, smaller wire used for theseapplications limits the total energy delivered to theequipment.

NEMP does not have the same amount of energyas a direct lightning strike, it doesn’t last as longand usually doesn’t generate the high peakcurrents (20kA) of lightning. Rise times are afunction of the coupling impedance to the unitunder test, as well as the unit’s dynamic surgeimpedance.

All of the grounding techniques for lightningprotection previously discussed apply to NEMP.Most importantly, long equipment interconnectionlines can couple NEMP energy. Long lines will alsohave larger voltage drops across them because ofthe faster rise times involved (L di/dt). Extensiveindividual shielding of the inter-equipment wiringis necessary. The most cost-effective way toprovide shielding for a room full of equipment is ascreen room enclosure. The next most importantpart is to provide a ground system that will actuallydissipate the fast NEMP pulse.


CHAPTER 10

Security Cameras, CATV,GPS & Satellite Protection

O utdoor Closed Circuit Television (CCTV)Security Cameras can be a prime targetfor lightning. They are usually mounted

on a building or other vantage point such as a metalo r wood pole. A lightning strike can not only

destroy the camera, but candamage your control console

due to energy flowing backthrough the coax andcamera power wiring.

When lightning strikes atower or other large

structure, there is a highpeak voltage at the strike

point flowing outward anddownward through any path it can

find to earth ground. A support pole develops ahigh L di/dt peak voltage drop along its length toearth ground. A large steel reinforced structurecan conduct the energy to earth ground throughits steel reinforced concrete footers and electricalground system. A camera mounted and groundedto a building with steel reinforced construction willusually have less inductance to ground than acamera mounted on a self-supported tower or pole.Less inductance to earth ground means less peakvoltage at the camera.

When lightning strikes a wood or other insulatingsupport, whatever voltage is necessary to continuethe arc is developed at the strike point to overcomethe resistance of the non-conducting structure.This usually has catastrophic results.

The same conditions exist for both examples. Ahigh peak voltage occurs at the strike point with

reference to earth ground. The video and powerwiring to the camera are insulated from the strikepoint by the electrical circuitry involved and theexternal covering around the wire. The energy willflow through the camera in an attempt to equalizethe wiring with the instantaneous peak voltageoccurring at the strike point.

To protect your equipment, you must provide alow inductance path to earth ground for lightningenergy and install properly rated protectors for allinterconnected wiring from the camera to theoperating console. A properly rated protector atthe camera allows the wiring to be equalized tothe peak voltage at the strike point without allowingdamaging voltages across the camera circuitry. Anappropriate protector at the console blocksdamaging voltages incoming from the camerawiring.

A camera mounted on a building should begrounded to the building’s structural steel as nearthe camera as possible. Use 1-1/2 inch copper strap.If the camera is mounted on a metal pole, it shouldbe grounded to the pole and a proper groundsystem installed at the base. When mounted on awood or other insulating support, the camerashould be grounded to a 1-1/2 inch copper straprunning from the camera mount to a proper groundsystem installed at the base. An additional 1-1/2inch copper strap would run from a lightning rodor diverter to the ground system at the base.Separate the two straps on opposite sides of thepole and connect together below grade. Sidemounting the camera or providing a diverter abovethe camera provides some additional protectionfrom a direct strike.


A proper ground system would be capable ofdispersing large amounts of lightning energy(usually electrons) into earth ground quickly. Thefaster it disperses electrons, the less time there isfor damaging surges to flow in the coax and powerwiring back toward your operating console.

The ground system under a metal pole could be acombination of a steel reinforced concrete base(Ufer Ground), radials and ground rods. If possible,exothermically weld a #2/0 AWG stranded wire tothe steel mesh before pouring concrete. Attachthis wire to a “J” bolt on top of the pad after thepole is erected. Use another wire welded to themesh to attach additional radials with ground rods.If the concrete base already exists, attach additionalradials with ground rods to any “J” bolt.

Be sure to remove paint and corrosion. Use adouble nut attachment with joint compound. Spaceadditional ground rods at least two times theirlength from each other and from the “UferGround.” (See illustration.)

When grounding a wood or insulated support, tietogether both 1-1/2 inch straps, below grade, to aradial strap and ground rod system. A good layoutfor a “rapid response” low resistance/inductanceground system would be four 8-foot ground rods,one at the base and three spaced 120 degreesand 16 feet out forming an equilateral trianglecentered on the base of the support. Each groundrod would directly connect with below grade 1-1/2inch straps to the rod under the pole. (Seeillustration.)

Protector type varies depending on camera powerrequirements and environment. Numerousconfigurations are possible.

For example: A camera powered by 120 Vac wouldrequire a PolyPhaser IS-PLDO-120-15A at thecamera. If there is insufficient space in theweatherproof housing, an IS-PSP-120 MOV/Gastube hardwired shunt protector at the camerapower input can be substituted.

If the camera is powered by 24 Vac, a PolyPhaserIS-SPTV twisted pair protector could be placed at

Ufer Ground

the camera and an IS-PSP-120 protector wiredacross the primary of the 120Vac to 24Vac powertransformer at the console end of the cable.

In both examples an IS-75BB (75 ohm BNC femaleconnectors) would be inserted in the video coaxat the camera and control console ends. Theprotectors should be in a weatherproof locationunless a water-tight version is ordered.

A “Rapid Response” low inductance ground system

Ground Rods “Sphere of Influence”

8-foot GoundRod at each

Point “ ”

16 feet from center to Ground Rods

8-foot Radius each “Sphere of Influence”


To Radial Ground System or Equipment Building Ground Loop

J Bolts

Rebar Cage


#2/0 To Rebar Connection

Tower Leg


DIVERTER

WOODEN UTILITY POLE

SECURITY CAMERA

POWER LINE PROTECTOR IS-PSP-120

GROUND DOWN LEADS ARE SOLID COPPER

8"-12" SEPARATION

COAX TO SWITCHER

TO SWITCHER GROUND SYSTEM AND OTHER GROUND RODS

GROUND RODS COPPER CLAD STEEL 5/8" DIA. 10' LONG

CADWELD OR CLAMP CONNECTION OF “GROUND DOWN LEADS”

COAX RUN TO CAMERA

3 OR MORE LOOPS IN COAX 12" DIA. TAPED TOGETHER LAY ON FLOOR

TO CENTRAL GROUND

SYSTEM (1-1/2" COPPER

STRAP)

POWER LINE PROTECTOR IS-PLDO-120US15A

IMPULSE SUPPRESSOR(LIGHTNING ARRESTOR)(WRAP CONNECTIONS FOR WATER TIGHTNESS)IS-75BB POWER

CORD

IMPULSE SUPPRESSOR IS-75BB

TO CABINET (1-1/2" COPPER STRAP)

IGA-90V PROVIDES GROUNDISOLATION TO 90VDC WHENUSED WITH IMPULSESUPPRESSOR IS-75BB

Some cameras use the coax cable for 24 Volts dcpower input and video output. The power isinserted at the control console end and “pickedoff” the coax in the camera.

A PolyPhaser Part # 093-0421W-A [Special] (75BB+26 Volts, <100 mA, dc-15MHz) could be insertedin the coax at the camera and control consoleends.

“Ground Loops” can occur whenever long videocoax runs are used. The usual symptoms includehorizontal black bars (hum bars) moving verticallythrough the picture. Ground Loops are createdwhen a potential difference exists betweengrounds. This potential difference can allow inducedpickup from power lines or current flow fromdissimilar ground potentials, i.e. Power Plants orSubstations.

The Ground Loop can be eliminated at the cameraby using an IS-75BB base band protector andinsulating the camera from local ground at the topof the pole. The isolated ground adapter does not

have a path to ground until a predetermined voltagefrom coax shield to local ground is exceeded andremains switched to ground until another lowervoltage condition exists. At the control console end,an IS-75BB protects the center conductor. Notethat this arrangement will only work for externallypowered dc or low voltage ac powered (through atransformer) cameras with no “safety” ground.

In all cases, the IS-75BB mounted on a “single point”ground plate should be used at the console endwith a few turns of coax to add series inductancebefore connecting to switcher or monitor. An IS-PLDO-120US15A in-line ac power protector is alsomounted on the ground plate with the controlconsole and any power supplies for remotecameras plugged in to it. (See illustration.) Theground plate should be connected with 1-1/2 inchstrap to an external low inductance ground system.Do not rely on the third wire ground in the ac wallsocket.


CATV CONSIDERATIONS

Almost everything discussed so far can be appliedto a CATV head end. Trunk line amplifiers are verysimilar to the tower top amplifiers alreadydescribed. They are powered via the coax line,usually with 60Vac (60Hz). The ac power isseparated from the RF, optimally protected, thenrecombined on the equipment side.

The power line (neutral ground at the utilityentrance) should be interconnected with the cabledrop ground. The interconnection should be a wireplaced in the ground (buried) for low inductance (ifthe soil is conductive) or should be a buried strap.Getting the ground connection in place andkeeping it there can be simplified if the cableentrance point is located close to the power groundlocation.

GPS LIGHTNING PROTECTIONAND ANTENNA PLACEMENT

The first consideration for a GPS antenna is a clearview of the sky, preferably 360 degrees. In theusual installation the GPS antenna is located low,close to the equipment building roof, or if anoutdoor cabinet, mounted on the cabinet or verylow on the adjacent monopole/tower. A directlightning hit to the above mounted antenna isunlikely. Mounting on an equipment building roofor cabinet is the safest place since the potentialrise on the outside of either of these structureswould be more or less equal with the potential onthe inside. The PolyPhaser protector is there toequalize the differential in potential that occursbetween center conductor and shield of the coaxcable on its way from the antenna to the receiver.

The zone of protection from various lightning rodtypes is an argued topic. Many claims are madefor different configurations. If a Franklin rod is below60 feet, we can assume a 45 degree “cone ofprotection”. If above, we should apply the “rollingball” theory (see pages 2-3).

If the GPS antenna is mounted on the monopole/tower, (since this is the structure we expect to behit) there will be an inductive voltage drop occurringduring the event that will be distributed down thestructure to earth ground. This voltage drop is theresult of the fast rise time lightning current pulsetraversing the inductance of the structure. (E = Ldi/dt). If the GPS antenna is mounted on thisstructure it will be elevated to a potential higherthan the equipment building or cabinet. There willbe current flow on the shield and center conductorof the coax cable towards the receiver. A coaxcable grounding kit or PolyPhaser integrated groundentry panel will direct the shield currents towardearth. A PolyPhaser coaxial protector will “turn on”and direct any current on the center conductortowards earth. Proper shield grounding and centerconductor protection are essential to receiversurvival.

Questions regarding GPS LNA protection in theantenna are valid but usually not considered in thisapplication. The antenna element at GPS frequencyis usually “grounded” and does not have thecapture area to couple much energy to the preampinput, the problem is with the output. In roof orcabinet mounting there is not the potential thatcould occur with a monopole/tower mount. If theGPS antenna support structure is elevated inpotential (due to its inductance), the GPS antenna/LNA will also be elevated to a potential determinedby the voltage distribution across the structure,and the height of the GPS antenna mounting onthe structure. Since the coax shield is usuallycommon with the GPS antenna mounting bracket,current will flow down the shield. The voltagedifferential at the top of the coax between theshield and the not-yet-elevated center conductorwill appear across the LNA output circuitry. TheLNA output could be destroyed in the attempt tobring the center conductor up to shield potential.If another protector were installed at the outputof the LNA, any voltage differential between centerconductor and shield would “turn on” the protector.Current flow that would have gone through theLNA output now goes through the protector. TheLNA would survive. The top protector could becombined with a voltage “pick-off” for power tothe LNA. There are protectors in this configuration.


There have not been many failures (that we knowabout) with LNA’s. The higher the GPS antenna ismounted on the support structure, the moreprobability of damage.

SATELLITE DISH CONSIDERATIONS

Most satellite dish antenna pier supports areencapsulated in concrete, which retains moisturefor 15 to 30 days after a rain or snow melt. Itabsorbs moisture quickly and yet gives up moisturevery slowly. Concrete’s moisture retention, itsmineral content (lime and others), and its inherentpH base of more than +7 means it has a readysupply of ions to conduct current. Concrete’s largevolume and great area of contact with thesurrounding soil allows good charge transfer tothe ground.

If a 4-inch pipe is placed four to five feet down inan 18-inch diameter (augered) hole and the hole isfilled with concrete, it will provide a good start fora satellite dish ground system. This type “ground”is referred to as a “Ufer,” so named after HerbertG. Ufer. (For more information on Ufer grounds,see Chapter 3.)

In areas of good soil conductivity (100 ohm-metersor better), the “Ufer” may be adequate for theantenna ground. Interconnect a below grade barecopper wire from the antenna pipe to the electrical(power) ground stake or rod. The copper wireshould be 10 gauge or larger, at least 8” or deeperin the ground, and should be exothermically weldedto the pipe. The weld will ensure a good mechanicaland electrical connection. Exothermic welding is asimple process which joins even dissimilar metalswithout a problem and requires a fixture (mold), or“one shot” connection kit.

If the water table at the installation site is over 10feet deep, the Ufer ground should be augmentedwith mechanically coupled pairs of 10-foot rodsplaced 20 feet deep and spaced 20 feet apart. Thefirst ground rod location can be at the antenna,and the second should be 20 feet away, in thedirection of the equipment building and connectedto the below grade bare copper wire going to thebuilding perimeter ground loop.

Additional radials should be used to augment theUfer grounding. RF and control cables for the systemshould follow the line of grounding rods which arespaced 20 feet apart (overhead view).

If the installation is in a rocky area and it is virtuallyimpossible to install ground rods, a radial systemmay be used. Grounding may be accomplished bylaying 10 or more lengths of 10 gauge or largerbare copper wire, at least 50 feet long in a radialfashion connected to and going out from theantenna base, much like the spokes of a wheel.

These may be laid on the surface although it ispreferable to bury the wire. The antenna base mustbe interconnected to the equipment buildingperimeter ground loop.

In sandy terrain (dry sand), increase the inter-connect wire size to #2 AWG or larger copper wire(copper strap may be used), and shorten thedistance between the rods to 10 feet.

It is always best to measure the soil conductivity,but it is not essential to strive for a 5 ohm groundsystem. By using the proper I/O (Input/Output)protectors, the equipment can survive moststrikes. A satellite dish system may have thefollowing I/Os: 1) 120Vac power; 2) RF coax cable;3) dc polarization control; 4) actuator/positionercable.

TV

RECEIVER / ACTUATOR

IS-PLTVRO-P*WALLOUTLET

IS-TVRO-P*GROUNDROD "H"

SIGNAL &CONTROLCABLES

IS-TVRO-P

POWERMAIN

PROTECTOR

UTILITYGROUND

GROUND RODSEVERY 20'

* EITHER HAVE AN IS-TVRO-P OR AN IS-TVPLTVRO-P.EITHER SHOULD BE GROUNDED TO GROUND ROD "H".

IF AN EXISTING POWER LINE PROTECTOR IS GROUNDED TO GROUND ROD "H", THEN AN IS-TVRO-P (OR THE SAME TYPE UNIT AS IS USED AT THE DISH) SHOULD BE MUNTED AT GROUND ROD "H", IN LIEU OF THE IS-PLTVRO-P UNIT INSIDE THE HOUSE.


The power line can be a two-way street. If alightning strike occurs away from the system butnear a utility pole, it can travel to the equipment.A direct strike to the satellite antenna will causecurrents incoming to the equipment, elevating itand exiting to the utility line, causing damage inthe process.

A power line protector that plugs into the wall hasthe equipment power cord series inductance fromthe equipment to it and also has an even longersafety wire ground run to the distribution panelbefore it gets to the utility ground rod. The fastrise time pulse of a lightning strike, it is not reallygrounded at all!

A power mains protector mounted directly to thebreaker box across the 240 volt mains to neutraland ground can protect the whole site fromincoming surges. A second protector mounted orgrounded directly to the equipment chassis will“switch” surge current incoming from the antennato the antenna system ground and to the perimeterground loop.

RF CABLE PROTECTION

Surge current will propagate on the coax creatinga differential in as little as 15 feet. The centerconductor energy differential can damageequipment, inside (receiver) and outside (LNA,LNB, down converter, etc.). A coax protectorclamping at the +22 volt level for LNB’s with lowloss and good VSWR over the 450 to 1450 Mhzrange is required. The specifications of theprotector will vary based on the type of systembeing installed.

POLARIZATION CONTROL &ACTUATOR

The polarization rotor or polarity switch is anothersource of surge current to the equipment. Inaddition, the actuator or driver must have aprotector for the controller and also the motor andswitches.

The system will need two sets of protectors, oneat the antenna, which is grounded to the antennaground system, and the other at the equipmentbuilding, grounded to the perimeter ground loop.

At this point, protection has been established forthe system I/Os, but in the event of a direct striketo the feed, damage can still occur to the LNA/LNB. To help stop direct strikes, take a 10-foot rodand pound it in no more than 6 feet away from theantenna. Couple a second 10-foot rod to extendat least 2 feet above the highest point on the dish.Interconnect the rod assembly to the antennaground system. This assembly should act as adiverter, sending up a streamer towards a steppedleader, and attracting the strike to itself.

77Lightning Protection & Grounding Solutions for Communication Sites

AppendixLightning Protection Myths and Traditions

Tables & Formulas

Definition of Terms

Bibliography

APPENDIX


Lightning Protection Myths& TraditionsTOWER LIGHTNING DOWNCONDUCTORS

The first myth is adding a lightning rod to a towerand a copper down conductor running the lengthof the tower and separately connecting to theground system or one ground rod. The concept isto divert the lightning energy around the tower.Usually the joint resistance between towersections, averages about 0.001 ohm and will createa voltage drop during a tower strike. Thedownconductor is to eliminate these resistivedrops. However, the strike’s inductive voltage drop,for either the copper cable or the tower, farexceeds the resistive voltage drops, and theground potential elevation at the tower base nearthe one isolated ground rod, makes it ineffective.The tower inductance is many times smaller thanthe down conductor’s inductance (due to the muchlarger surface area or skin effect, even with theferrous effects of the tower steel). The voltagedifference between the down conductor,attempting to carry the strike with its large L di/dtvoltage drop, and the tower will be furtherincreased by the tower induced opposite EMFvoltage on the down conductor. This huge voltagedifference will be so great it could arc 1.5 feet!Unless the tower and downconductor areseparated by this impractical distance, they will beconnected by the arc’s conductivity, and currentwill still traverse the tower. The coax cable on thetower has a much larger circumference/lowerinductance than the downconductor and is muchmore effective at “shunting” the tower sections(if that were a concern). The coax cable is alsogrounded to the tower at the top, middle, andbottom.

The proximity of a bare copper downconductor tothe tower’s vertical legs can be detrimental. Rain

water coming into contact with copper can carryions that react with the galvanized coating on thetower, causing it to wash off. (Natural rain water--not acid rain--has a pH of +5.5 to 6.0, which isacidic.) Acids will attack copper and with time eat itaway. This can decrease the life of the tower (andits safety) as the tower rusts.

Tradition is hard to overcome. In the early daysof broadcasting, when medium wave (AM) stationswere beginning to operate high-poweredtransmitters, it was noted that, due to high rfcurrents on the tower, small arcs were occurringbetween corroded joints on existing (old) towers.Arcing noise was effecting the transmitted audio.When a continuous downconductor was placedfrom top to bottom of the tower, the transmitternoise was greatly reduced or went away.

Ship to shore radio communications on low tomedium wave frequencies sometimes usedinsulated (from earth) towers for both transmittingand receiving. In some cases the transmit towerand the receiver tower were (in terms ofwavelengths) close together. Sincecommunications were on several frequencies atthe same time, it was noted there was receivesignal degradation that could not be removed byreceiver filtering (then). When the same corrodedjoints on the receiver tower were bypassed by adownconductor, the degrading “noise” was greatlyreduced or went away (again). Now radioengineers, being an observant group, noted theimprovement and carried the downconductor ideaover to new tower site construction. However, the“tower bypass conductor” only shunted the towersections and did not go to earth ground--it couldnot with an insulated tower.


Since large downconductors on structural lightningrod installations were standard, and since agrounded tower appears to be a very large lightningrod, “common sense” suggested a large down-conductor on the tower with perhaps a lightningrod on top. The lightning rod would send an“upward going streamer” to connect with thedownward step leader from the cloud and “protect”the antenna from a direct strike and subsequentphysical damage. This is a good idea if the “lightningrod” is higher than the antenna (and won’t interferewith the antenna’s radiated pattern). A largedownconductor, insulated from the tower, wouldthen be routed down in the tower to a separateground rod or grid. Perhaps, to further isolate thelightning energy from the tower, the lightning rodwould be insulated from the tower. Do you think a4” insulator or an insulated wire would “isolate” atower from lightning current after the “bolt” hasjumped two miles to hit you?

Note: If a grounded communications tower werein a multiple frequency, high rf environment, therecould be a slight chance of rectification or re-radiation of combined rf energy that might causereceiver degradation or transmitted spurious. Adownconductor could be of some value if the towerjoints were questionable.

GROUND RADIAL CIRCLES

Myth: Connect equally spaced ground rods (fromthe tower base) together to form a circle orconcentric circles.

“Radial wires should be interconnected by wirerings.”

If this grounding system were installed at a sitewith a poor conductivity earth surface layer, thelarge inductance of the interconnections to therings will make the rings ineffective. A ground ringwill launch electrons mostly outside itscircumference. Two closely spaced rings are liketwo ground rods attempting to couple charge intothe same volume of earth. The electrons from eachrod are repelled, reducing the effectiveness of theinvestment in buried copper.

Tradition: “Once again let us return to thosegolden days of yesteryear” and revisit ourbroadcast engineer friends. Early MW (AM)broadcast stations were concerned with daytimegroundwave propagation to adequately serve theirlocal audience. The radiated “take-off angle” of theirtransmitted signal had to be as low as possible, sothe grounding system under the tower had tocapacitively couple to the earth body efficiently.Experimentation to determine optimum groundingsystem geometry determined that multiple radialsoutbound from the tower base closelyapproximated the ideal solid plate capacitivecoupling to earth.

As experiments continued, an ideal groundingsystem model started to emerge. During theexperimental phase, our engineers noted thatgrounding systems approaching the ideal modelappeared to lessen lightning damage to transmittingequipment. Radials every 3 degrees (120)extending out the height of the tower, or ¼ waveof the operating frequency were determined asoptimum. Experiments with concentric wire “rings”with radials were conducted to more closelyapproximate the ideal flat capacitor plate.

Once again, this observed decline in lightningdamage, due to “radials and rings” was carried overto new tower construction. Newer computer-generated tower and grounding models haveimproved on the original concept and matched thegrounding system more closely with coverage area,ground conductivity, and terrain requirements. Butground rings remain as a traditional method evenon non-radiating grounded towers. The real answeris multiple radials with ground rods (see chapter2).

NOTHING WILL SAVE YOU

Nothing will save the equipment if the tower takesa direct strike. This myth came from those whodid not install a proper grounding system. It isdifficult to allocate additional funds for the materialsand labor needed to include a good groundingsystem when a new site is under development.Often the lowest bidder gets the job erecting thetower. Unless a specified grounding plan is written


in the bid package, the bare minimum is quoted orgrounding is omitted altogether. Those who thinkthat three eight-foot rods around the tower baseis a good grounding system are on the right trackonly if the tower will be erected on the beach athigh tide.

Even with planning, no one can say if an installationwill or will not survive, since the intensity of a hit isa variable. It is safe to say that less damage (ifany) will be sustained with a properly groundedtower.

TO GROUND OR NOT TO GROUND

One of the best sayings is really not a myth at all.“A grounded tower is more likely to be hit.” As wesaw in the graph in chapter one, the taller the objectthe more frequently it’s struck. Why ground a towerif it means that it is more apt to be hit? Look at itfrom the reverse point of view--a tower that isn’tgrounded and is hit will send energy out anyavailable path towards a lower potential. Theequipment could be the path of choice. If the toweris properly grounded, the owner has control. Thewhole concept of lightning protection is to controland direct the lightning surge energy so it doesthe least amount of harm or damage.

ELECTROSTATIC DISSIPATIONDEVICES

Some people think if you have many sharp pointsaround a tall object, it will give off enough ions toenshroud the object and eliminate lightning strikes.If this were true, towers with many antennas, ortrees in a forest, would never be hit! Trees arenot a good conductor. But with the very highimpedance of static E fields, they do conduct, sincelightning does hit them. Whether due to amountain, building, or tower height, strikeoccurrence goes up as the square of height overaverage terrain. Brushes or static dissipationdevices will not prevent a strike.

One of the selling “points” for wire brushes is theyreduce the electrostatic voltage between thecloud’s charge center and the site. This is indicated

as being done with a grounding system thatreduces the induced electrostatic voltage in theground by the brush creating ions, blown away bythe wind. This would be a great idea, but the groundcan provide the induced voltage faster than thebrush can create ions. Imagine the site as a threedimensional object in free space. The site’s surfacearea has an ability to store a charge. The brush (ora few sharp points) can create a current, in an Efield of 10kV/m, of about 0.5A. This is a 20k ohmload. The R×C time constant of this discharge,compared to the site ground (typically 5 ohms), is(20k/0.5 = 4000) 4000 times faster! This meansthe earth (with a 5 Ohm grounding system) canprovide electrostatically induced voltage to thetower 4000 times faster than the brush candissipate it!

Wind blown ions will not cancel the earth/cloudcharge and prevent a strike from happening. If nowind is present the ion space charge will driftupwards. The ions make the air more conductiveand can help direct the stepped leader (precursorto the strike) to the brush resulting in an upwardgoing streamer (return stroke). Brushes will notstop the formation of upward going streamers,although they may slightly delay it. They may helpstop corona from a sharp object under labconditions, but under real site conditions they arenot effective. We simulated a strike to a dissipationbrush with our largest surge generator and foundthat when struck, the brush explodes in spray ofmolten steel creating a fire hazard for any structurebelow.

Those who sell this kind of product claim manyhappy customers with no equipment damage. Onlythose who do not have a good grounding systemwould buy such devices. When they are installed,a good grounding system is always included withthe installation. The good grounding system is whatreduces damage! A US Navy report, done incooperation with the USAF, NOAA, and the FAA,states that after several years of study, dissipationdevices have not prevented any lightning strikes(If they worked, NASA would be using themextensively). The FAA did another study and hasvideo tapes showing them being hit (again).


The myth we want to hear: “We don’t know if allthe money we spent for lightning protection wasworth it. We haven’t been hit yet!” How does oneknow that a strike has not occurred? (PolyPhaserhas a line of strike counters.) Undamagedequipment is not a true indicator. It only takes onestrike. If lightning protection was installed, then“no news is good news.”


Water & Soil Resistivities

Type of Water or Soil Resistivity in OhmsWater in oceans 0.1 - 5Ground water, well and spring water 10 - 150Lake and river water 100 - 400Rain water 800 - 1,300Commercial distilled water 1,000 - 4,000Chemically clean water 250,000+

Clay 25 - 70Sandy clay 40 - 300Peat, marsh soil and cultivated soil 50 - 250Sand 1,000 - 3,000Glacier 3,000 - 10,000


FormulasMULTI-CHANNEL COMBINING

Combining consists of summing multiple channelstogether to feed onto a common cable. Lowfrequencies and higher frequencies will in timereach a peak voltage together of the same polarity.This voltage summation peak can have more peakpower than the sum of their RMS powers. Sincegas tube protectors are voltage sensitive, theymust be designed to handle combiner outputs orturn-on RF voltage peaks can result. This can causemajor problems at other frequencies (intermod,spurs, interference, etc.). Each protector designedfor combiners is listed as to its total voltage peak(VT) that it can handle without RF turn-on. This totalis the summation of all the voltage peaks for eachchannel being combined.

VT = V

P1 + V

P2 + V

P3 ... + V

Pn

where VP = 1.414 • X • eP

ch • 50

(Pch

= channel power out of combiner)

and for a VSWR X =1.1 to 1 1.051.2 to 1 1.091.3 to 1 1.131.4 to 1 1.171.5 to 1 1.21.86 to 1 1.3

CALCULATING INDUCTANCE:

Copper Wire:L(in µH) = 0.508l [2.303 log

10 (4l/d) – 0.75] x 10–2

Copper Strap:L(in µH) = 0.508l [2.303 log

10 (2 l/(w + t)) + 0.5

+ 0.2235 (w + t)/l] x 10–2

Where: l = length in inchesd = diameter in inchesw = width in inchest = thickness in inches

IS-158E & 318E HIGH POWERGAS TUBE SELECTION:

VP = 1.414 • X • eP

ch • 50

VP

= Voltage peakP

ch= Power per channel

X = VSWR (same as Multi-Channel Combining)


Definition of Terms

AMPEREAn ampere (current) is a (coulomb/second).

BANDWIDTHDifference in frequency between the upper andlower 3dB down response frequencies.

BI-PHASEFound as a power feed to most U.S. homes. Derivedfrom a center tapped transformer, it contains twohot phases (180°) with a center tap neutral return.Normally supplied as two 120 volt single phaseswith 240 volts available across both phases. Theneutral return is usually earth grounded.

CAPACITANCEMeasured at 1.0kHz unless otherwise stated.

CLAMPTo clip. To hold turn-on voltage as current isincreased. Turn-on voltage is the same, or nearlythe same, as “on” voltage drop.

CLAMPING RATIOThe ratio of voltage drop at a given current to theturn-on voltage.

CLAMPING SPEEDMeasured with full lead length using a 1kV/nswaveform in a 50Ω system, with >300MHz orlarger bandwidth.

COMBINERThe summation of multiple transmitters into onetransmission line. The peak voltage from eachsignal will be additive and will be higher than thesum of the power would indicate.

COMMON-MODEPertaining to signals or signal componentsreferenced to ground.

COULOMBMeasurement of charge. Often used to indicatethe amount of transferred charge through a gastube to determine gas tube life. “Q” abbreviation.A coulomb is (current x time).

CROWBARTo turn-on and clamp close to ground level. Havinga high turn-on trigger voltage and a low “on”voltage.

DIFFERENTIAL MODEReferenced only between conductors (notreferenced to ground).

DIPLEXER(TV Broadcasting). The combining of twotransmitters into one transmission line. TV visualand aural.

DUPLEXERSimultaneous receive and transmit on onetransmission line. Where a T connector splits/combines the signals to two groups of filters. Thereceiver filter passes the receive frequency whilerejecting (band stop) the transmitter’s frequency.The transmitter filter passes its frequency whileattenuating the Class C transmit noise at the receivefrequency.

EMI/RFIElectro Magnetic Interference/Radio FrequencyInterference. Broad spectrum noise or interferingsignals.

EMPElectro Magnetic Pulse, usually referred to as theman-made generation by detonation of a nuclearbomb at a high altitude, which generates a veryfast pulse (RF) which can be captured by antennasand long unshielded lines. Sometimes referred toas NEMP, HEMP, etc. Lightning can also generatean EMP near the event. Referred to as LEMP.

EMP RATEDRated as having a fast enough turn-on time orfiltering to protect against the effects of an EMPevent.


FARADAY SHIELDAn electrostatic (E field) shield made up of aconductive or partially conductive material or grid.A Faraday cage or screen room is effective forprotecting inside equipment from outside radiatedRF energies.

FILTERING (EMI/RFI)Measured in a 50Ω system — loaded. As per MIL-STD-220.

FREQUENCY RANGEThe bandwidth over which both the listed maximumVSWR and Insertion Loss specifications are valid.

GROUND IMPEDANCEThe ground resistance and the inductance/capacitance value of the grounding system. Alsocalled dynamic surge ground impedance.

GROUND LOOPAn undesired potential EMI condition formed whentwo or more pieces of equipment are inter-connected and earthed for shock safety hazardprevention purposes.

GROUND RESISTANCEThe resistance value of a given ground rod orgrounding system as measured, usually by a fall ofpotential (3 stake) method, using a 100Hz signalsource.

HFHigh Frequency – normally from 3 to 30MHz.

HOUSED USE ONLYFor indoor use, or must be further enclosed orrain-proofed for outdoor usage.

IMPEDANCENominal impedance of the device. The variation ofthis impedance with frequency is measured asVSWR.

IN-LINEPower or signal passage through unit. In serieswith line. Usually a multi-stage protector. Bestprotection method.

INSERTION LOSSLoss of a device across the stated frequencyrange. This type of loss is due to the insertion ofthe unit in series with a signal path.

JOULESA unit of energy. One joule for one second is equalto one watt of power. Joules is (current x time xvoltage).

LEAKAGE CURRENTUsually measured at 50 or 60Hz with 120, 240 or480 volts ac. However, it can be ac or dc at a specificvoltage and frequency.

LOOP RESISTANCETotal resistance as measured across the input withthe output shorted.

MAXIMUM PEAK LET-THROUGH VOLTAGEMeasured at a given surge current using a givenwaveform, and using >300MHz bandwidth acrossa 50Ω impedance. (Note: this 50Ω impedance maybe dc blocked [large bandwidth compared to thesurge frequencies present] and 50Ω resistive load[termination]).

MAXIMUM POWERMaximum Continuous Wave (CW) transmit power,without unit degradation.

MAXIMUM SURGEThe maximum single surge current and specifiedwaveform that can be handled by a device withoutfailure during the conduction of that waveform andwhich ends the life of the device for conductingsuccessive waveforms, but does not allow anygeneration of outward projectiles.

MULTI-STRIKE CAPABILITYIn most applications current sharing will occur, andin a direct strike event the unit will survive to workagain.

POWERPower is (voltage x current) or a (coulomb/second).


RECEIVER MULTICOUPLERSometimes with an amplifier, this device has oneantenna line and multiple outlets.

RFRadio Frequencies — any and all frequencies thatcan be radiated as an electromagnetic wave (planewave).

SAFETY GROUNDThe local earth ground. The earth ground whichgrounds the neutral return. The wire may be greenor bare and can be through a metal conduit. It maybe earth grounded as many times as needed.(Neutral must only be grounded once at the entrylocation).

SHFSuper High Frequency – from 3000MHz to 30GHz.

SHUNT PROTECTORLine-to-ground. No power or signal passagethrough unit. Not in-series with line.

SINGLE PHASEA true single phase supply. Usually a two-wiresystem with one hot phase and a neutral return. Asafety earth ground is also present.

SKIN EFFECTThe gradient conduction and propagation of RF orRF components of a surge on the outer surfacesof conductors.

TEMPERATUREThe extremes of operating or storage that the unitor unit parts have been tested to under MIL-STD-202 for thermal shock.

THREE PHASEIt consists of sinusoids 120° apart on at least threewires (Delta) and often four wires (Wye). The fourthwire is a grounded neutral return. In a Delta systemthere is no reference to ground and thus it is moresusceptible to lightning problems.

THROUGHPUT ENERGYThe total energy that will be let through the deviceusing the indicated surge waveform.

TOTAL SURGE ENERGYTotal sum of surge energy for all lines of a protectorunit. Measured in joules. The minimum total energywhich results in the failure of the unit.

TRANSFER IMPEDANCEReferring to coax, is the impedance to transferinto or outside the coax at various frequenciesusually below 1MHz. Due to loss of skin effectattenuation or shielding at these low frequencies,coax can be susceptible to interference and noiseas well as the radiation of such signals.

TURN-ON TIME – GAS TUBEThe amount of time that exists in the period thatoccurs when the ramp voltage barely exceeds theturn-on voltage of the device, and the point atwhich 50% of the peak voltage is achieved duringthe turn-on (crowbar) process. Measured in a 50Ωsystem with >300MHz bandwidth.

TURN-ON VacThe maximum ac sine wave voltage that can bepassed with the peaks just at the turn-on Vdc level.

TURN-ON VdcTurn-on voltage at 1mA dc with a ramp of 100V/mstypical.

UHFUltra-High Frequency – normally from 300 to3000MHz, however in this catalog we breakout800 to 1000MHz separately even though it is withinthis category.

VHFVery High Frequency – from 30 to 300MHz.

VLFVery Low Frequency – from 300Hz to 3kHz.

VOLTA volt is a (joule/coulomb).

VSWRVoltage Standing Wave Ratio (VSWR) of the deviceacross the stated frequency range. VSWR is theamount of reflected signal due to an impedancemismatch.

VT MAXThe max peak voltage of all combined waveforms.V

total is used for multi-coupled or combined transmit

signals.


Bibliography & FurtherReading(1) Decible Products Inc. About Lightning

(2) Block, Roger R. The “Grounds” for Lightning &EMP Protection.

(3) Eaton, J.R. “Impulse Characteristics of ElectricalConnections to the Earth.” General Electric Review,October 1944.

(4) Fagan, E.J. and Lee, R.H. “The Use of ConcreteEncased Reinforcing Rods As GroundingElectrodes. IEEE Transactions On IGA, July/August1970.

(5) Freeman and Sachs. ElectromagneticCompatibility Design Guide. Sachs/Freeman Assoc.Inc., 1981.

(6) IEEE Transactions On Broadcasting. Vol. BC-25, No 1, March 1979.

(7) IEEE/ANSI C62.41 – 1991 RecommendedPractice on Surge Voltages in Low-Voltage AC PowerCircuits

(8) Lightning Protection, Code NFPA 780.

(9) Military Handbook 419A, Grounding, Bonding,and Shielding for Electronic Equipment and Facilities

(10) Schure, Alexander. Electrostatics, New York:John F. Rider, June 1958.

(11) Towne. Lightning Arrestor Grounds. G.E.Review, 1932.

(12) Ufer, H.G. “Investigation and Testing of FootingType Grounding Electrodes for ElectricalInstallations” IEEE Transactions Paper #63-1505,Power Apparatus Systems Vol. 83 pp 1042-1048:Oct. 1964.

(13) Uman, Martin A. “All About Lightning” and“Lightning,” New York: Dover Publications,1986,1984.

(14) Viemeister, Peter E. The Lightning Book.Cambridge: The MIT Press, April 1972.

(15) Weiner, Paul. “A Comparison of ConcreteEncased Grounding Electrodes to Driven GroundRods”. IEEE I&CPS Conference. May 1969.

Please check out our web site(www.PolyPhaser.com) for informative links onlightning.


NOTES

lightning protection grounding solutions for communication guide.pdf · lightning protection &...

Documents